Mozilla weighs in on India’s draft data protection bill

Yesterday, on July 27th, 2018, the Justice Srikrishna Committee of Experts, set up by the Government of India, made public its final report and the draft of India’s first comprehensive data protection law. We have long argued that the enactment of a baseline data protection law should be a national policy priority for India, and we’re pleased to see India take an important step forward towards enacting real privacy protections.

The legislation is groundbreaking in several respects, codifying principles and enforcement mechanisms that Mozilla has advocated are foundational to a robust data protection framework. But the law is not without loopholes, many of which threaten to dislodge these strong foundations.

Mozilla Chairwoman Mitchell Baker observed: “India’s data protection law will shape the relationship between users and the companies and government entities they entrust with their data. This draft bill is a strong start, but to truly protect the privacy of all Indians, we can’t afford loopholes such as the bill’s broad exceptions for government use of data and data localization requirements. Mozilla will continue to advocate for changes; with this bill, India has the opportunity to be a model to the world.”

As this bill makes its way to law, an open and consultative process is essential. We will continue to advocate to the Government to make necessary changes in the bill.

Top level highlights from the bill include:

  1. Obligations – Strong obligations that apply to both private companies and the government, including purpose limitation, collection limitation, data security, documentation, and a general duty to process data in a way that’s “fair and reasonable” and “respects the privacy” of the person. This law applies to Indian residents’ data wherever it may be processed.
  2. The Data Protection Authority – Creation of an independent Data Protection Authority with expansive powers including investigatory, adjudicatory, and punitive powers, as well as a separate Adjudicating Officer to take complaints, impose penalties, and mete out compensation to individuals. However, the independence of the adjudicatory authority and appellate tribunal responsible for legal proceedings related to data protection violations is severely lacking. The qualifications and nominations of those serving in these bodies are entirely prescribed by the government, as are the procedures of the bodies themselves. The system as it currently stands has far too much delegated authority to the Central Government. The power of setting qualifications and procedures and nominating individuals to serve in the adjudicatory authority and appellate tribunal should be reserved for the DPA, which operates independently of the government.
  3. High standard for consent – For consent to be valid it must be free, informed, specific, clear, and capable of being withdrawn. This sets a high bar for companies seeking to validate their actions on the basis of consent. “Explicit consent” is required for processing of sensitive data.
  4. Grounds for Processing – The bill allows for data processing for “reasonable purposes”. While similar in intent to the GDPR’s “legitimate interest” ground, the bill limits the potential for abuse by providing conditions on the basis of which data may be processed, as well as an illustrative list of categories that fulfil these conditions. We think this is an improvement on the GDPR standard, which as we noted in our submission, can “easily be abused by companies” who may argue that “innovation” itself is always a reasonable pursuit, even where it may put the privacy of users at risk.
  5. Biometric Data – Biometric data and the Aadhaar identification number are included in the definition of sensitive personal data which comes with stricter obligations. The bill includes a generally inclusive and progressive list of sensitive personal data including data related to religious or political belief, sexuality, transgender, and intersex status. Section 106 bars processing certain forms of biometric data as determined by the Central Government, unless the processing is explicitly permitted by law. This provision could be used to curtail the lax limitations on the handling of Aadhaar data.
  6. Individual Rights – Individuals are provided comprehensive rights of correction, updation, and data portability. However, rights to deletion and to object to processing, which are guaranteed by other data protection laws around the world including the EU’s GDPR, are notably missing. Users may have to pay for certain rights, which could entrench existing inequalities and create haves and have-nots for privacy.
  7. Data Processing for Security – Data processing for security, intelligence, and law enforcement purposes must be “necessary and proportionate”, and must be authorised by a law passed by Parliament. While a quick reading of this bill might look like there are exceptions for “security of state” data processing and the potential for mass surveillance, Section 42.1 actually provides substantive protections. For the number of intelligence and security agencies that currently operate in a legal vacuum, this bill would necessitate regulation, and one that meets the standards of “necessary and proportionate”. The “necessary and proportionate” standard is a critical part of international human rights law around surveillance, as well as the Puttaswamy judgement, and prevents this bill from ushering in mass surveillance. Section 42.1, if enacted, will necessitate a public debate about the appropriate limits of Indian government surveillance — data processing for security, intelligence, and law enforcement purposes cannot happen in the absence of such a debate and subsequent law.
  8. Cross-border Data Transfer – Cross-border data transfer is made possible through a variety of means, but rejects consent alone as sufficient for transfer, and conditions transfers on having a high level of data protection in place.

Some Particularly Worrying Provisions

  1. Data Localization – A copy of all personal data is required to be stored in India. As we have argued, data localization is bad for business, users, and security. Notwithstanding the protections on processing in the interest of the security of the state, it’s hard to see that this provision is anything but a proxy for enabling surveillance.
  2. Government Data Processing – A large swathe of government data processing activities for both sensitive and non-sensitive data, including for the provision of any service or benefit to a data principal, is exempt from the requirement of obtaining consent. However, the government needs to show that any processing of personal data is “necessary” and processing of sensitive personal data is “strictly necessary” for the exercise of any function of the State authorised by law for the provision of service or benefit. This means that the government must prove that processing data such as workplace, address, or phone number is “necessary” and processing data such as passwords, financial data, and biometric data is “strictly necessary” for any function that would provide a service or benefit. There is no necessity standard for government processing of non-personal data.