BSAResponseof White Paper Data Portection Framework India
BSAResponseof White Paper Data Portection Framework India
BSAResponseof White Paper Data Portection Framework India
BSA | The Software Alliance is the leading advocate for the global software industry. 1
Our member companies are at the forefront of data-driven innovation, and they have a deep
and longstanding commitment to protecting the privacy of personal information. BSA
appreciates the opportunity to provide comments on the White Paper of the Committee of
Experts on a Data Protection Framework for India (“White Paper”). As a global organization,
BSA actively follows privacy developments around the world. We have consistently
highlighted that an effective privacy regime provides appropriate protections for individuals’
personal data while also spurring innovation that is fueling the global economy.
Our comments below reinforce this view and respond to a number of specific issues
raised in the White Paper. Among other things, the comments emphasize the following key
points:
• The law should differentiate between and clearly define data controllers and
processors;
• The law should provide appropriate, flexible bases for processing data;
• The law should facilitate cross-border data flows and avoid burdensome
restrictions on data transfers, such as data localisation requirements;
1 BSA’s members include: Adobe, Amazon Web Services, ANSYS, Apple, Autodesk, AVEVA, Bentley Systems,
CA Technologies, Cisco, CNC/Mastercam, DataStax, DocuSign, IBM, Intel, Microsoft, Oracle, salesforce.com,
SAS Institute, Siemens PLM Software, Splunk, Symantec, The MathWorks, Trend Micro, Trimble Solutions
Corporation, and Workday.
Page 2
BSA recommends limiting the scope of India’s data protection law or framework to
entities or activities that have a sufficiently close connection to India. Specifically, BSA
recommends that the Government limit application of the data protection law to data
processing performed by an individual or legal entity, whether public or private, provided that:
(1) Indian residents are specifically targeted; (2) the personal data that is the object of the
processing is purposefully collected from data subjects in India at the time of the collection,
and (3) such collection is performed by an entity established in India through a stable
arrangement giving rise to a real and effective level of activity, or subject to India law by virtue
of international public law. Under this standard, the mere accessibility of a website in India or
the use of a language used in India would be insufficient, on their own, to establish the
applicability of India’s prospective data protection law.
In addition, BSA believes that the criteria suggested above to govern the applicability
of India’s data protection law will ensure effective enforcement of orders against foreign
entities.
Consistent with data protection laws in many other jurisdictions, BSA recommends
that the protections in India’s data protection law apply only to natural persons. This limitation
would tie India’s data protection law to the rights and interests of individuals, which are the
ultimate source and justification for the data protection principles upon which the White Paper
builds. As the White Paper highlights, the right to privacy that the Supreme Court of India
recognized in Puttaswamy vs. Union of India is derived from the right to life and personal
liberty guaranteed under Article 21 of the Constitution of India, as well as the Constitutional
guarantees of autonomy and dignity of an individual. In other words, data protection
principles protect interests that are tied to personhood; it is not clear that they extend
coherently to juristic persons. Further, extending data protection law to juristic persons is not
necessary to protect their interests in information and would create significant uncertainty
about how to apply the law. This uncertainty could chill data-driven innovation and harm
economic growth. Accordingly, consistent with the Expert Committee’s provisional view, the
law’s scope should be limited to natural persons.
In addition, it is critical for the Government to make any new data protection law
purely prospective and to provide a reasonable period of time between the enactment of a
data protection law and its effective date. Individuals, businesses, and the Government will
benefit more from an orderly transition than with one that is abrupt and requires catch-up
under threat of enforcement. Although BSA does not, at this time, suggest a specific
transition period, we note that, in other settings, legislators have allowed a two-year transition
period. Below, we offer several factors to consider in connection with determining the
appropriate transition period.
Many BSA member companies conduct business internationally, and any data
protection law that India enacts will become part of a complex global regulatory landscape.
Companies will need to continue to comply with their existing obligations under the data
protection laws of other countries and regions. To make their compliance efficient and
provide the best experience for consumers and customers, businesses will likely strive to
implement one set of data practices across all of their operations, including those covered by
India’s law. Undertaking compliance in this environment is a complex endeavor.
Page 3
Of course, companies will also need to ensure that they comply with any new or
unique requirements that India’s law creates. Companies may need to invest considerable
time and resources in modifying their data systems and practices. Along the same lines,
companies currently use several different mechanisms to transfer data around the world –
e.g., binding corporate rules, contractual clauses, APEC Cross-Border Privacy Rules
(CBPR)– and they will need sufficient time to update any contracts or arrangements to take
into account any new obligations.
Finally, the global data protection landscape may become more complicated while
India develops its law. For example, the European Union’s General Data Protection
Regulation (GDPR) will go into effect on 25 May 2018, and the EU is currently considering a
further change to its data protection requirements through the proposed ePrivacy Regulation.
These developments will add to the complexity of businesses’ compliance obligations and
should be taken into account in determining the appropriate transition period. Moreover, it
may be helpful for the Government to assess the effectiveness of the GDPR’s
implementation to determine whether any lessons learned could be useful guidance as India
develops and implements its own data protection framework.
BSA recommends referring in the law to “personal data,” as this term has been
adopted in several other data protection laws and frameworks, such as the GDPR and the
Organisation of Economic Cooperation and Development’s (OECD) Guidelines on the
Protection of Privacy and Transborder Flows of Personal Data.
More important than the label is the scope of data that the law covers. As indicated in
our response to Part II, Chapter 2, the law should apply only to natural persons. Beyond that,
the law should protect information that, if mishandled, would have a meaningful impact on an
individual’s privacy. Accordingly, the law’s definition of “personal data” should be limited to
data that is reasonably linked to an identified or identifiable individual.
If the scope is not limited, and stringent legal obligations apply to a broad range of
data regardless of its context and the risk of harm to specific individuals, the law is likely to
have a chilling effect on everything from cybersecurity to improved customer services without
having an actual benefit on personal privacy.
BSA also recommends excluding anonymized data from coverage of the data
protection law. A suggested definition of anonymised data is “data that is not reasonably
linkable to a specific individual.” By definition, such data is not personal and therefore does
not implicate individual privacy interests.
Clearly excluding anonymised data from the definition of “personal data” – and the
data protection law as a whole – would benefit individuals and the economy. The exclusion
would give businesses the incentive to develop and use anonymisation techniques, thereby
reducing privacy and security risks. At the same time, the ability to use anonymised data
outside the framework of a data protection law will encourage innovative uses of data.
Finally, we note that, within the scope of information covered under the definition of
personal data, the law should distinguish between fully identified personal data and
pseudonymous data, relaxing appropriate requirements when data is processed in
pseudonymized form. Pseudonymous data is organized according to a randomly generated
identifier that is not used in any other dataset and from which other information that could be
used to connect information to a specific individual has been removed. Because such
techniques mitigate the risk that third parties will link the information to a specific individual,
pseudonymised data should be subject to less stringent rules than personal data in which
Page 4
BSA recognizes that cultural considerations may counsel in favor of creating certain
sensitive data categories, but we also recommend that the Government take a rigorous, risk-
based approach to its determinations. Defining certain types of personal data to be “sensitive”
may be consistent with a risk-based approach, but care must be taken to limit such
categorization. Other data protection laws have limited that categorization to racial or ethnic
origin, political opinions, religious beliefs, genetic or biometric data, health information, or
information regarding sexual orientation. Even if these categories of information are
designated as sensitive, BSA recommends that the Government avoid imposing restrictions
solely based on these categories. Rather, BSA recommends a risk-based approach to
determine whether heightened protections should apply in these limited circumstances based
on the context in which data is used.
In Puttuswamy, the Supreme Court of India also appears to recognize the importance
of a risk-based approach in defining the right to privacy. In pages 201-203, the judgment
describes various aspects of privacy by identifying four zones (i.e. from ‘personal’ to ‘public’)
and the potential risk to two aspects of freedom, i.e., the freedom to be left alone, and the
freedom for self-development.
Examining risks arising from the use of data, by contrast, would direct companies to
focus their efforts on reducing risks in the context of their own data processing operations.
This focus will lead to greater flexibility in data processing and promote innovative uses of
data while also providing appropriate data protections.
BSA’s overarching view is that India’s data protection framework should be flexible
and risk-based. Personal data processing typically depends on a set of distinct but
interrelated operations. For example, the ability to use personal data may require transmittal,
storage, and retrieval. In practice, companies rarely design data protection programs that
focus on atomic operations in data processing. Instead, BSA members and most modern
enterprises take a holistic approach toward risk management and compliance with applicable
data protection laws. Covering all facets of data processing need not be inconsistent with this
goal and may, in fact, reduce the risk that the law will unintentionally exclude or establish
differential treatment for certain actors.
The law should differentiate between and clearly define “data controller” and “data
processor.” Differentiating between data controllers and data processors serves to establish a
clear allocation of roles and responsibilities and helps to clarify complex cases, where the
data is processed by more than one entity (e.g., outsourcing of processing).
• "Data controller" means a natural person, juristic person, public authority, agency or
other body which, alone or jointly with others, determines the purposes for which and
the manner in which any personal data are, or are to be, processed
• “Personal data processor” means a natural person, juristic person, public authority,
agency, or any other body which processes personal data on behalf of the data
controller
The primary obligation for ensuring compliance with the applicable personal data
protection law should fall on the data controller. The data controller retains a direct
relationship with the data subject, determines what information to collect and for what
Page 6
purposes, decides how it is used, with whom it is shared and under what terms. By contrast,
the data processor acts on behalf of the data controller and does not make the essential
decisions affecting compliance with core data protection obligations. Accordingly, a data
processor’s main obligation should be to follow the instructions of the controller and ensure
the security of the personal data it processes.
Contracts between data processors and data controllers are the most effective
means for governing processor responsibilities with respect to personal data. Importantly,
controllers and processors should have the flexibility to negotiate their own contractual terms,
without mandatory, prescriptive language provided by the law.
The clear allocation of liability among controllers and processors is critical and
prevents confusion from arising in the complex system of relationships that underlie modern
data processing operations. Maintaining this allocation would also avoid disturbing the
existing economic and contractual relationships between processors and controllers. Finally,
making data controllers primarily responsible for ensuring compliance with data protection
law comports with common sense, because legal authorities and individuals know to whom to
turn to in case of a problem.
In short, data controllers should have the primary obligation for ensuring compliance
with applicable privacy law, while data processors should only be required through
contractual mechanisms to comply with data controller instructions, assist the controllers in
meeting their own compliance obligations, for example, by providing controllers with the
means to comply with data subject rights with respect to the data processors handle, and to
ensure the security of the data they process.
Chapter 7 – Exemptions
BSA supports including an exemption for data processing that is conducted for
research, historical, or statistical purposes. Personal data can provide extremely valuable
insights into a broad range of social and economic issues, aiding research and supporting
advances that enable improvements in health, education, transportation, and other areas.
The law should provide that data must be kept secure against unauthorized use or disclosure
during research, historical, or statistical processing, but data protection law should not
otherwise hinder use of the exemption. BSA can identify no reason to restrict or prohibit the
use of a research/historical/statistical exemption when research is subsequently published or
used for a commercial purpose. The exemption should focus on whether further processing
of personal data will serve a bona fide research, historical, or statistical purpose at the time
that the processing is to be conducted and on the merits of the proposed use.
BSA supports the ability of law enforcement organizations to access digital evidence
in connection with lawful criminal investigations, provided there is sufficient privacy
protection. It is therefore appropriate that exemptions to data protection law enable law
enforcement investigators to access data under specific and controlled circumstances, as
outlined below. With regard to potential exemptions for national security or public safety,
legitimate threats to national security or public safety will constitute activity covered under
criminal law, and investigations seeking access to data for such purposes would therefore
likely be able to do so through an appropriately scoped exemption for criminal investigations.
Accordingly, BSA recommends developing a law enforcement/national security exemption
that focuses on criminal investigations.
(3) requires prior judicial approval for access to content data; and
(4) does not seek to mandate exceptional technical access to protected data (such
as encrypted data).
Law enforcement organizations should be able to access data held outside the
country’s sovereign territory only through established bilateral and multilateral mechanisms
governing such access, such as Mutual Legal Assistance Treaties (MLATs) that establish
safeguards for the transfer and processing of personal data to ensure privacy protections.
Page 8
Under a potential law enforcement exemption, individuals should have the right to
access data that has been collected about them, to know the purposes for which such data
has been collected, and to know by whom it has been collected. Except under exceptional
circumstances, private entities holding personal data should have the right to inform
customers when they have been directed to provide personal data about the customers to
law enforcement organizations.
The need to transfer data across borders arises in a variety of contexts. Global data
flows enable multinational companies to scale global operations, startups to use cloud
services to obtain digital infrastructure at lower costs, and small and medium-sized
enterprises to use digital platforms to find customers and suppliers abroad. As one example,
Lazada, Southeast Asia’s most popular online shopping platform, operates local-language
online retail stores in Indonesia, Malaysia, Philippines, Singapore, Thailand, and Vietnam.
Lazada used cloud-based services to rapidly expand its product range by streamlining the
vendor registration process, reducing the sign-up and onboarding time from two months to
ten days.
not give access to the entire data set. In addition, access to real-time global data enables
companies to detect and address critical threats. For example, one leading cybersecurity
company with operations in more than 50 countries searches for matches in its global
database of emerging threats, often blocking the spread of a new attack.
Notably, data analytics, relying on cross-border data flows, have also contributed
significantly to the social good, including enhancing public health and safety. For example,
researchers around the world leverage data analytics to respond to natural disasters,
including in the wake of the 2015 earthquake in Nepal, where they conducted a real-time
analysis of mobile phone patterns to assist in disaster relief efforts.
In short, the economic and societal impact of cross-border data flows is substantial,
and any proposal for a data protection framework should aim to preserve the ability to
transfer data globally.
To protect individual privacy while “harness[ing] the benefits of the digital economy,”
(White Paper at 4), BSA recommends that the law explicitly recognize the ability to transfer
personal data outside of India. However, the law should also recognize that no single data
transfer mechanism alone is likely to meet the needs of modern technologies and services.
While a data protection law may helpfully recognize certain data transfer mechanisms, it
should also allow companies to adopt alternative, legally binding protections.
Accountability should provide the touchstone for data transfers. The accountability
model, first established by the OECD and subsequently endorsed and integrated in a variety
of data protection frameworks around the world, provides an approach to cross-border data
governance that effectively protects the individual and fosters streamlined, robust data flows.
Canada, for example, has implemented this principle effectively into its national data
protection law, avoiding preemptive restrictions on data transfers while at the same time
ensuring that organizations remain responsible for personal information in their possession or
custody that is subsequently transferred to third parties for processing.
Indeed, the accountability model requires organizations that collect personal data to
be responsible for its protection, no matter where or by whom it is processed. As such, any
organizations transferring personal data must take steps to ensure that any obligations — in
law, guidance, or commitments made in privacy policies — will be met.
Further, as part of ensuring the free flow of data, the law should prohibit data
localisation requirements for both the public and private sectors. As discussed in the
response to Chapter 9, such measures can frustrate efforts to implement security measures,
impede business innovation, and limit services available to consumers.
BSA also urges the Government to refrain from imposing an adequacy requirement
on international data transfers. An adequacy requirement is unnecessary under an
accountability-based framework because of organizations’ ongoing obligation to ensure
Page 10
appropriate processing. As such, accountability should remain the focus of data protection
law.
Moreover, experience with the adequacy requirement in the EU’s Data Protection
Directive (which is also included in the GDPR) points to several significant drawbacks in
instituting such a requirement. Adequacy reviews are time-consuming and burdensome,
resulting in few adequacy findings. Adequacy reviews also focus heavily on formal legal
standards, which may not provide a complete picture of the safeguards in a country.
Conversely, the existence of formal legal protections in a country does not necessarily
guarantee a high level of legal protection in practice. Finally, adequacy could limit the ability
of companies in India to provide modern technologies and services. In short, an adequacy
requirement is unlikely to add significantly to the level of data protection in India, but
adequacy would likely hinder innovation and growth in India’s digital economy.
Some established mechanisms for transferring data, including the EU-U.S. Privacy
Shield and Standard Contractual Clauses in the EU, have particular rules related to sensitive
data, but the transfer of that data should be broadly permitted. A risk-based approach should
be used to determine whether any additional safeguards should be imposed on the
processing and transfer of sensitive data, as discussed in our response to Part III, Chapter 6.
However, BSA strongly recommends against imposing bans on the transfer of certain
categories of data, including based on the sensitivity of data.
We understand that the Ministry of Home Affairs has developed manuals to classify
data based on various factors, including sensitivity of the information, impact on India’s
national security, and strategic or administrative considerations. BSA submits that any such
data classification exercise should be based on the principles of transparency, predictability,
and fairness and should be a clearly defined, publicly available, narrowly tailored, and easily
administrable exception to the general rules governing international transfers. Otherwise, the
prohibition could create uncertainty that slows the growth of Digital India.
As referenced in the response to Chapter 8, BSA does not recommend the adoption
of data localisation requirements. This approach is consistent with developments in
international trade agreements, which generally prohibit localisation requirements as a
condition for doing business in a country. These agreements include limited exceptions for
localisation measures necessary to achieve legitimate public policy objectives, provided they
are no broader than necessary to achieve the objective and are not applied in a manner
constituting a means of arbitrary or unjustifiable discrimination. Although important concerns
about data protection, security, and law enforcement have led some countries to impose
localisation requirements, they typically do not provide an effective means of addressing the
underlying concerns and have significant negative consequences, including increasing costs
Page 11
to economies and local companies, limiting access to global services, and hindering
innovation.
As the White Paper recognizes, data localisation measures have had a negative
economic impact on GDP of several countries. (White Paper at 70.) For example, existing
data localisation measures already have had an estimated -.8% impact on GDP in India, and
an estimated 1.8% in Vietnam. It has also affected exports for several countries, resulting in a
-1.7% export loss in both Indonesia and China. According to Gartner, public cloud services in
India are projected to grow at 38% this year to total $1.81 billion. Data localisation
requirements could impede this growth, as they often impose significant costs on the
countries that adopt them.
Further, data localisation may prevent companies from offering services within a
country, because it may be too costly or otherwise impractical to do so. Individuals have
fewer and more expensive choices as a result. The White Paper cites studies showing that
companies would be required to pay 30-60% more for their computing needs than if they
could go outside the country’s borders.
In addition, companies that are subject to data localisation requirements may not
have unfettered access to innovations in other countries. This will hamper the development of
domestic innovation that would otherwise occur by enabling local companies to access
foreign cloud service platforms, analytics tools, and application programming interfaces
(APIs). Further, data localisation can reduce the size of local data sets, restricting the value
that can be gained from cross-jurisdictional data consolidation and analysis. As a result,
domestic companies would face higher costs, slower innovation, and isolation from
customers in other markets. These effects could be most pronounced for India’s start-up
companies, many of which depend on access to low-cost cloud storage and computing
services. As the White Paper acknowledges, many Indian start-ups forego making large
capital investments in computer hardware and leverage cloud services to meet their business
needs. (White Paper at 72-73.) In short, imposition of data localisation requirements would
be contrary to the goals of promoting a “Digital India.”
The White Paper highlights the wide range of existing laws in India – including in the
financial, information technology, telecommunications, and health sectors – that implicate the
processing of data. We encourage the Government of India to ensure that any general data
protection framework it develops supersedes any other legal requirements that apply to the
protection of personal data so that a consistent approach can be developed that establishes
clear rules, which will provide legal certainty for businesses of all sizes and help to facilitate
compliance.
More generally, in light of the complexities of global compliance, the law should be
applied in a manner that is consistent with international best practices to facilitate the ability
of companies to provide products and services to Indian residents.
Page 12
Chapter 1 – Consent
(Questions 1-5)
Consent is an important basis for handling personal data. However, BSA urges the
Government to recognize other legal bases for processing personal data, as discussed in our
response to Part III, Chapter 4 of the White Paper. These additional bases for processing
include the legitimate interest of companies handling the data, the performance of contracts
with the data subject, and compliance with legal obligations. The data protection framework
need not identify a primary ground for processing. Instead, legal grounds should be generally
applicable, and it should be up to the data controller to determine the relevant ground(s) –
and to ensure that its processing activities comport with such grounds.
The standard for determining the level of consent that is appropriate should be
contextual. In circumstances that do not implicate heightened sensitivity, implied consent may
be appropriate. In today’s world, a large amount of data is created through individuals’
interactions with Internet-connected devices, and express consent is not suitable or practical
in all instances.
As the White Paper points out, it “is a questionable assumption” whether individuals
can fully engage in privacy self-management through informed and rational decisions about
data collection and use, given the vast number of instances of data collection in the modern
world. A data protection framework that relies too heavily on individual consent will
overwhelm individuals with decisions and render them unable to identify truly consequential
decisions. This can lead to two further negative consequences: (1) stymying growth and
innovation in the digital economy; and (2) not meeting consumer privacy expectations by
leading consumers to “click fatigue,” where users simply accept whatever terms are
presented to them without fully reviewing or understanding the information presented to them.
Recognition of implied consent can relieve some of this burden. The law could
recognize implied consent for data processing associated with certain, specified purposes,
such as fraud prevention and service fulfillment. The deployment of more data-intensive
services in the future should also be allowed to rely on implied consent to collect personal
data for processing that is not harmful, is reasonably expected, and allows services to
operate. For instance, a public transportation service that uses electronic fare cards should
not be required to obtain express consent each time an individual uses his or her card.
Concerns about imbalances in bargaining power are less relevant in these kinds of cases,
because the data processing poses little or no risk and directly benefits individuals.
Chapter 3 – Notice
(Questions 1-7)
Consistent with our views on consent, BSA recommends that notice requirements
reflect the context of data processing. When express consent is required, a clear and
conspicuous notice that provides individuals with information relevant to their choice is
appropriate. In other situations, a requirement to provide prominent notice is unlikely to
advance privacy protections. Requirements that result in over-notification may result in
individuals ignoring notices that are important for an express choice that they need to make.
Also consistent with the idea that notice should support choices that are contextual,
the law should not prescribe the form or content of a privacy notice. Individuals are using an
increasing number of devices from which personal data may be collected, and the displays
on these devices are becoming smaller or even disappearing. Finding ways to provide
effective notice in connection with such a diversity of devices is challenging, and prescriptive
requirements in legislation are unlikely to address these challenges. Addressing this ongoing
challenge should involve the input and efforts of industry, the Government, and other
stakeholders. Because the issue of notice in a more connected, data-intensive world is
unresolved, BSA cautions against prescribing specific means of presenting notice, such as a
“consent dashboard.” The specific requirements for this approach are not well understood. In
the absence of prescriptive requirements, however, we do recognize that dashboards could
be useful tools to promote user control.
While the drawbacks of presenting privacy notices separately from a device (i.e.,
“decoupling” notice from the device) are real, this presentation may be the only means of
presenting full and accurate information in a public document. Presenting privacy notices in
this manner also affords the space and flexibility to offer translations in multiple languages, to
reflect the diversity of individuals who use a product or service.
In addition, privacy impact assessments as proposed in the White Paper under this
Chapter are not relevant in the context of notice and choice. Such assessments are
conducted by organizations on their own processes to see how they might implicate the
privacy of the individuals whose data they hold, collect or process. They are not applicable in
the discussion relating to notification to individuals.
Finally, BSA does not support the use of a “data trust score” to encourage the
development of effective notices. As presented in the White Paper, the data trust score
concept only provokes further questions about how the score should be defined and what
body should be responsible for providing it. It is unlikely that the quality of a notice – let alone
a broader set of data practices – can be meaningfully boiled down to a single score. The
effort and resources that would be required to develop a data trust score would divert
resources of the government and businesses that would be better spent developing more
substantive data protections.
Page 14
BSA supports providing multiple legal grounds for processing personal data. Consent
is an important basis to offer, but it should not be the only legal basis for processing. As
discussed below, other grounds of processing may provide a better fit for certain data
processing situations. Consequently, these grounds may direct companies to consider
requirements that are less burdensome for individuals and companies, while also resulting in
more effective data protections.
Legitimate Interest
The legitimate interest legal basis for processing personal data, which is incorporated
into many privacy laws around the world, is vital for enabling companies to conduct data
processing that is necessary to carry out their business operations. Legitimate interest serves
a particularly important role where it may not be suitable or practicable to obtain consent, or
where it is premature to enter into a contract with the data subject. For example, businesses
may need to handle personal data as part of network security or fraud prevention efforts.
Seeking consent for such activities is not only impractical but may lead to harm to a
significant number of individuals.
In addition, the law should recognize that companies may need to address user
conduct that directly harms their interests. For example, if a financial institution is seeking to
recover an outstanding debt and needs to collect, use, process, and/or disclose personal
data as part of the debt-collection process (e.g., to debt-collecting agencies), it may not be
suitable to request the data subject’s consent to do so, but there is a legitimate interest that
would justify the handling of the personal data.
BSA recognizes that putting legitimate interest into practice not only requires the
specification of additional factors to consider in determining whether it applies to an instance
of data processing but also the exercise of judgment in each application. There are multiple
ways to address this challenge. For example, the data protection framework itself could
outline factors, based on further input and dialogue as the Government moves forward.
In the EU, for example, processing is permissible without the need for consent when
it is necessary for the purposes of the legitimate interests pursued by the controller, except
where those interests are outweighed by the impact of the processing on the rights of the
data subject. The EU’s Article 29 Working Party has outlined several factors for companies to
consider in weighing data subjects’ interests and applying this ground for processing.
The White Paper queries whether it is appropriate for data controllers to conduct the
balancing test of individual rights that has been associated with the legitimate interest ground
for processing. We note, however, that the data controller is obligated to assess and
remediate data protection risks in a variety of circumstances under data protection laws, and
they can exercise that judgment capably under these circumstances too. Several countries
around the world have also recognized legitimate interest as an important basis for data
transfers, including other non-EU countries within Europe, as well as South Africa. Other
countries, such as Brazil, the Philippines, and Ghana, are considering legislation that would
also recognize a legitimate interest ground for processing.
individual participation rights, will help ensure that data processing based on legitimate
interest respects individuals’ privacy rights.
Companies should also be able to handle personal data to comply with legal
obligations. Businesses are subject to a wide range of legal obligations, including financial
reporting rules, other regulatory requirements, and obligations arising from court proceedings.
In some instances, companies must handle personal data to satisfy these legal obligations.
For example, under Indian law, any organization that is affected by a cyber security incident
must report the incident to the Indian Computer Emergency Response Team, “as early as
possible,” However, in the event of a large-scale incident affecting several users, it may not
be practicable for the organization to obtain the consent of each individual user before
complying with this legal obligation. The absence of any alternative grounds for processing
would also hinder the ability of the organization and the relevant authorities to take swift
corrective action. Therefore, any privacy framework should ensure that companies can
continue to comply with these requirements.
Contractual Performance
Other Bases
In addition to the foregoing examples, there are several other potential circumstances
that could serve as valid legal bases for handling personal data. These could include
performing tasks in the public interest and protecting the vital interests of individuals, which
are also recognized under other data protection laws. We recommend that the Government
adopt a flexible approach that both protects individuals’ privacy and preserves companies’
ability to carry out their legitimate business operations and provide innovative services.
BSA concurs with the White Paper’s assessment that the “advent of newer
technologies such as Big Data, data analytics and the Internet of Things may challenge the
relevance of the purpose limitation, as it currently exists.” (White Paper at 106.) Adopting a
strict version of purpose specification and use limitation, which would require data to be
collected for a specified purpose and used for no other purposes unless notice is given (and
perhaps additional consent is obtained), could impose undue constraints on data processing
and inhibit data-driven innovation.
As an alternative, BSA suggests permitting use for purposes that are compatible or
related to the purpose for which data was collected, a standard that is also reflected in the
APEC Privacy Framework, would be more consistent with new technologies and modern
data-driven services. We note that several other elements of a data protection framework
would provide ample protection under this modified purpose specification and use limitation
principle. Accountable organizations will have processes in place to determine whether a
reasonable relationship between initial and subsequent purposes exists, and they would
document their decisions. Data security obligations would continue to apply to personal data
that is used for related purposes. In addition, individual participation rights in the data
protection framework, potentially including rights of confirmation, access, and rectification,
would continue to apply. Finally, processing for additional purposes would require a
Page 16
satisfactory legal basis under the data protection law. Taken together, these elements of the
data protection framework would provide robust safeguards for individuals and cause
companies to make careful decisions about whether to process data for additional purposes.
Assessment of Compatibility
The data controller is in the best position to determine whether a subsequent use of
data is reasonably related to the initial purpose. The criteria governing compatibility deserve
more discussion as the Government develops its data protection framework, but the
considerations should focus on context and individual expectations. For example, the law
could require data controllers to consider whether the additional processing extends the initial
purposes of collection and whether additional processing implicates greater sensitivity or risk.
Data controllers might need to consider their relationship with individuals whose data is at
issue and the circumstances under which the data was collected. For example, a company
that registers users who use one of its services might wish to use the contact information that
it collects in connection with the registration to notify users that additional services are
available. Such a use of registration data is compatible with the purpose of initial collection
and the relationship between the individual and the company.
BSA also suggests that the accountability-based outlook provides support for this
kind of flexible approach. Accountable organizations will have appropriate processes in place
to ensure that the organizations make and document well-reasoned decisions about whether
additional purposes are reasonably related to the initial purpose of processing.
BSA recommends that the data protection law provide general standards to govern
purpose specification and use limitation. This structure will provide a coherent framework to
put these principles into practice. Many companies provide services in different industry
sectors, and a general standard would enable them to achieve a more consistent application
of the data protection framework. Such consistency can be accomplished by consolidating
data protection oversight in one central regulator with broad understanding of different
industry sectors, rather than bifurcating the regulation amongst several regulators.
BSA recognizes that cultural and socio-economic considerations may counsel that
certain types of data be deemed sensitive, and that those categories, as defined in other data
protection laws, often include racial or ethnic origin, political opinions, religious beliefs,
genetic or biometric data, health information, or information regarding sexual orientation.
However, even if the Government designates these categories of information as sensitive,
BSA supports a risk-based approach to determine whether heightened protections should
apply to sensitive data.
Fundamental to the risk-based view is the notion that not all personal data is the
same, nor are all data processing activities the same, in terms of the risk that they pose to
individuals. BSA suggests that meaningful distinctions among personal data types must begin
with an identification of concrete, specific, and measurable harms. This harm-based outlook
will help keep sensitive data classifications, and any relevant processing restrictions, confined
to cases in which the processing of sensitive personal data poses a potential to cause
significant harm to individuals.
concrete, specific, measurable harm to individuals. An effort to determine whether such data
is sensitive is likely to go astray. Examining risks arising from the use of data, by contrast,
would direct the focus of companies toward reducing risk in the context of their own data
processing operations. For example, certain types of information implicate clear privacy
interests to the individual. However, it is still appropriate to apply a risk-based approach. For
example, if sensitive personal data is used by an employer to process employee benefits to
comply with relevant laws, the risk of harm is lower to the individual than it otherwise would
be in the context of a public disclosure and, therefore, should not be subject to the same
requirements.
More generally, this risk-based focus will lead to greater flexibility in data processing
and promote innovative uses of data while also providing appropriate data protections. Risk-
based data protections will be more effective in identifying and mitigating new or emerging
risks than will static processing requirements (or restrictions) based on data sensitivity.
BSA supports a storage limitation standard that permits companies to store personal
data for as long as it is relevant to the purposes for which it was collected. A specific,
quantitative limitation on the data retention period is not workable because of the vast
diversity in the data processing operations that are necessary to support modern data-driven
services. For example, any attempt by the Government to prescribe storage limitation
standards, based on duration, manner and format, as provided for under Section 67C of the
Information Technology Act, 2000, would be difficult to implement given the wide variety of
services and data categories that would fall under its ambit. Similarly, a qualitative standard
that is too strict will inevitably require companies to delete or anonymize personal data before
it provides its full value to individuals.
Other data protection mechanisms are better suited to ensuring that companies do
not retain personal data that is no longer relevant. Companies identify (or refrain from
collecting) data that is not relevant to their processing operations through their accountability
programs. Data security programs help to protect all personal data under a company’s control
from unauthorized disclosure and use. Anonymization also protects data from misuse and
misappropriation.
BSA’s proposal to permit personal data storage for as long as data is relevant to the
purposes for which it was collected avoids the pitfall of a specific, one-size-fits-all rule for
deletion or anonymisation. This standard would give companies the flexibility to provide
innovative data-driven services while ensuring that they remain responsible for deleting or
anonymising personal data that is no longer relevant. Other data protection processes,
including ensuring that data collected is relevant to specific purposes and maintaining data
security programs, also provide effective safeguards.
It is critical to strike the right balance between the scope of confirmation, access, and
rectification rights with the potential legal effects of data processing. The legal effects of data
processing vary widely with context. Some data processing leads to significant decisions
(e.g., employment and credit) by persons other than the individual, which can have a
significant impact on the individual. Other instances of data processing might only affect or
Page 18
influence choices that individuals make, such as which item to purchase online. In these
contexts, the individual remains the decision-maker. These two contexts raise substantially
different issues with respect to individual participation, and confirmation, access, and
rectification rights should take these differences into account.
Moreover, setting a fixed timeline implies that all needs behind confirmation, access,
and rectification are equal. As discussed above, they are not. Considerations for setting a
standard to respond to individual requests should take into account the impact of data
processing on the individual as well as whether the individual is the relevant decision-maker.
We also note that requiring a reasonable fee would be consistent with approaches in other
countries, and it would also enable resources to be spent on more specific, individualized
attention to these requests.
BSA strongly supports giving individuals more control over their data. At the same
time, any additional individual rights under a data protection framework should be consistent
with the goal of harnessing the benefits of the digital economy. (See White Paper at 4.) To
that end, it is essential that the law establishes requirements that are technically and
commercially reasonable to implement and does not prescribe how they must be
implemented.
Restriction of Processing
and 135.) BSA acknowledges that various issues could arise in the context of automated
decision-making. However, because this is an area where technology continues to evolve, we
emphasize that prescriptive rules could both frustrate innovation and fail to achieve the aim of
creating a framework that is “practically enforceable and effective.” We note further that this
continues to be an area of considerable research, including on how these emerging
technologies could be used to detect and address bias.
Second, the right to object should be limited to decisions based solely on automated
decision-making that can have significant legal or economic consequences, and where the
other rights and obligations in the framework are insufficient to guarantee protection of the
individual’s interests. Otherwise, a right to object to automated decisions could override the
structure of rights and obligations established elsewhere in the data protection framework.
Limiting a right to object to decisions with significant legal and economic consequences, such
as access to employment, housing, or credit, rather than expanding it to include areas that
simply provide information for consumers to make informed decisions, such as advertising,
would both carefully tailor it to instances in which there is a substantial risk of harm and
enable the application of objective criteria to ensure that the right is practically enforceable.
BSA supports the aim of giving individuals choice and control over their data. The
consent and individual participation rights, the storage limitation obligations, and other rights
and obligations discussed in the White Paper can help achieve this objective. We understand
that the Karanataka High Court, in Sri Vasunathan v. The Registrar General, recognized a
“right to be forgotten” in connection with highly sensitive information, and that the Supreme
Court of India, in the Puttaswamy case, further acknowledged how the ability to delete
information online can impact privacy. As the Government assesses how to incorporate these
concepts into a data protection framework, BSA outlines below how such a right might be
applied and identifies three important factors to consider in shaping deletion rights—clearly
defining the scope of any such rights, ensuring the balancing of other rights, and creating a
practical framework that is reasonable and practicable.
As the White Paper recognizes, courts in some jurisdictions, including the EU and
Japan, have applied what is commonly known as the “right to be forgotten” with respect to
online search results. BSA appreciates the concerns of protecting individuals’ privacy online.
We also recognize that unqualified extensions of this right, particularly if defined and applied
inconsistently across multiple jurisdictions, could hamper efforts to develop pragmatic,
effective solutions to protect privacy and negatively affect the larger Internet ecosystem.
Therefore, certain options – such as the possibility of extending such a right into the physical
space, or extending the right beyond delinking of online information – must be assessed in a
balanced and practicable manner, as emphasized by the Supreme Court in Puttuswamy.
search engines, since this is the context in which courts have articulated the right and is most
closely tied to the individual interest in ensuring that certain types of information remain
obscure. We recommend that the available remedy should be limited to blocking of certain
results that appear upon a search of the requester’s name, but the underlying link and
content should still otherwise be searchable. This would strike a more appropriate balance
among varied individual rights and interests.
More generally, we also highlight three key considerations for defining deletion rights.
First, the right should be carefully tailored to circumstances where the personal data
is no longer necessary for the purpose it was collected, the data subject has withdrawn
consent and no other legal ground exists to the process the data, the data pertains to a child
under the age of 13, or the data has been unlawfully processed.
Second, as the White Paper recognizes, it is vital that any requests to delete
personal data be balanced with other important rights, such as freedom of expression. In
addition, the law should also recognize other circumstances where the exercise of such a
right may not be appropriate, such as compliance with a legal obligation or defense of legal
claims, for public interest reasons, including public health, or for scientific or historical
research purposes.
Finally, any obligations associated with this right should require only that data
controllers make technologically and commercially reasonable efforts to delete data that is
implicated by an individual’s deletion request.
BSA supports the use of the principle of accountability in the Government’s data
protection framework, provided that the focus of accountability in practice is on the outcomes
of data processing, rather than prescriptive ex ante requirements and administrative
processes that add little to the objective of increased data protection.
BSA opposes strict liability standards in data protection laws. In particular, imposing
civil penalties under a strict liability standard would create excessive deterrence, leading to
less innovation and higher costs for products and services. Moreover, imposing direct, joint,
or several liabilities or other obligations on data processors would have a range of unintended
consequences, would undermine the relationship between these actors, and would create an
unjustified compliance burden. In addition, this could also have a negative impact on potential
investments in data processing and outsourcing services. Contracts between data processors
and data controllers are sufficient to ensure that data processors comply with data controller
instructions and to ensure the security of the data they process.
Finally, BSA urges the Government to refrain from adopting any insurance
requirements in connection with liability under the data protection law. Clarity in the law about
the allocation of liability and the range of potential monetary liability will provide the basis for
companies to make their own decisions about the need for insurance.
The APEC CBPR system provides an example of the successful public-private sector
collaboration to develop of a data protection code of practice. Although APEC economies led
CBPR development, they drew significant input and assistance from industry. The result is a
voluntary, accountability-based set of program requirements that facilitate privacy-protecting
cross-border data flows and are enforceable in APEC economies. The object of the CBPR –
facilitating data flows among a group of large, dynamic economies – illustrates the kind of
value that can justify the commitment of resources necessary to produce a code of practice.
In India, codes of practice have been used in different sectors, including the media
and advertising sector. For example, the “Code for Self-Regulation in Advertising” developed
by the Advertising Standards Council of India (ASCI) serves as the principal framework for
advertisers to demonstrate compliance with existing advertisement laws. The ASCI code
functions as a voluntary, self-regulatory mechanism, and it is recognized as a successful
industry-driven effort that is currently being used in India. BSA recommends that a similar
approach may be followed in developing suitable codes of practice in the context of a data
protection framework for India.
Moreover, the data protection framework should recognize that not all personal data
breaches represent equal threats. In many instances, the breaches pose no actual risks to
the individuals whose personal data was affected. Requiring notification only when there is a
material risk of harm focuses on breaches that individuals need to know about in order to
protect themselves. A breach involving information that permits access to financial accounts
is perhaps the clearest example of material risk of harm to individuals. A focus on breaches
with material risk also provides clarity as to the information that should be included in
individual notifications. For example, information about affected financial accounts and
specific steps that individuals can take to secure their accounts and prevent misuse of data
exposed in the breach would be important to communicate.
The material risk standard would largely prevent the issuance of immaterial notices,
principally by ensuring that notification is only required where there is a material risk of
identity theft or economic loss to the user. Notification in the absence of such risks may
create “notification fatigue,” leading to undue inconvenience for individuals as well as the
possibility that individuals will fail to take appropriate action in response to notifications that
indicate a real risk of harm.
Page 23
Furthermore, a personal data breach notification system should also exclude from the
notification obligation all instances in which the personal data in question has been rendered
unusable, unreadable or indecipherable to an unauthorized third party through any practice or
method that is widely accepted as effective industry practices or industry standards (e.g.,
encryption).
BSA sees little to gain from establishing different categories for data controllers
based on, for example, the harm that they may cause to individuals through data processing.
BSA supports a risk-based data protection framework, and risk should guide both the content
of companies’ obligations under the law as well as their decisions about how to implement
the law’s requirements. Attention to risk, combined with processes that provide organizational
accountability, would require companies to engage in an ongoing process of assessing and
mitigating their data processing risks. Categorizing data controllers would add little to these
efforts.
Moreover, the creation of data controller categories could create unintended, adverse
consequences for individuals and the economy. Many companies in the modern, data-driven
economy handle a wide diversity of personal data and engage in a broad range of data
processing. Requiring data controllers to fit within specific categories could disrupt the
efficiencies of having these different operations under a single set of data systems and data
protection processes. There is also a risk that data controller categories would be under-
inclusive or over-inclusive. This, in turn, could lead to under-protection of data, or an unduly
high level of protection. All of these outcomes would reduce companies’ flexibility and could
impede innovation while providing no data protection benefit.
Registration
BSA is opposed to any registration requirement for data controllers. Identifying and
addressing risks based on the nature or volume of personal data that a controller processes
should be integral to the controller’s data protection processes. The framework discussed in
the White Paper also makes consideration of risk key to the elements of a data protection law
that the Paper discusses. Creating a registration requirement is unlikely to add meaningfully
to companies’ awareness of risk. Instead, registration will mainly serve to create additional
compliance costs for companies as well as burdens for the regulatory or enforcement
authority.
processing, as discussed in the White Paper (at p. 168) may be a useful tool within a set of
accountability processes. However, making this assessment mandatory is unnecessary and
likely to lead to additional prescriptions about the form and content of the assessment, all of
which will add to the compliance burden and reduce flexibility.
Adding a legal requirement to conduct audits would impose significant additional cost
and inflexibility while creating few, if any, benefits for data protection. Similarly, granting a
data protection authority the legal authority to conduct audits would disrupt companies’
operations and divert resources from areas in which they add more meaningfully to data
protection.
BSA would welcome the opportunity to engage further on the details of adjudication
processes as more specifics of the data protection framework are developed. However, at
this stage, we point out that there are significant downsides to requiring enforcement
authorities to become involved in resolving individual disputes. This level of involvement
requires a large commitment of resources, which might be better spent investigating
significant or large-scale violations of data protection law. In the United States, for example,
the Federal Trade Commission does not resolve individual consumer complaints, focusing
instead on strategically selected enforcement actions that address harm to large numbers of
consumers, shape industry norms, or stop especially harmful practices.
Chapter 4A – Penalties
(Questions 1-6, 11)
Fair and reasonable civil penalties can add credibility to data protection law and deter
violations. Therefore, civil penalties are a potentially useful data protection enforcement tool.
However, civil penalties should be available only for willful or manifestly negligent
violations of legal requirements. Data protection enforcement should focus on accountable
practices, increased detection, and the promotion of feedback or guidance to improve
operational practices. Imposing civil penalties under a strict liability standard would create
excessive deterrence, leading to less innovation and higher costs for products and services.
Civil penalties also should not become the primary enforcement objective, nor should they be
used to threaten data controllers and data processors that act in an accountable manner.
The law should provide additional safeguards in connection with civil penalties. Most
importantly, the law should not link civil penalties to a percentage of a company’s worldwide
turnover. Using worldwide turnover to calculate a civil penalty has the effect of imposing a
global sanction, which in turn may punish a group of companies for the actions of a single
subsidiary. This type of penalty is at odds with fair and reasonable enforcement and could
deter further investment in India.
In addition, the law should require the data protection authority to consider the
circumstances – and mitigating factors – surrounding each alleged violation. Specifically, if
alternative remedies, such as imposing additional requirements on a specific company’s
processing activities following a full investigation, would address the violation, then the law
should direct the data protection to use that remedy.
Finally, the level of civil penalties should be tied to any additional programs or
processes that a company puts in place after a violation. These remedies should be designed
to prevent future violations and therefore reduce the need to rely on civil penalties for
deterrence.
Chapter 4C – Offences
(Question 1)
BSA opposes the creation of criminal liability for violations of data protection law. The
substantive requirements of data protection law, combined with monetary relief and conduct
remedies provided through administrative or civil judicial processes, are sufficient to protect
individuals’ privacy interests.