By Mishi Choudhary and Eben Moglen
On June 9, IBM through a letter by its CEO Arvind Krishna announced that it’s exiting the facial recognition technology business altogether and wouldn’t condone use of any technology used for mass surveillance and racial profiling that violate “basic human rights and freedoms”. This started a positive trend amongst technology companies. Microsoft president Brad Smith confirmed on June 11 that Microsoft won’t sell to US police its facial recognition tech either. Amazon also announced a one-year moratorium on selling law enforcement agencies access to “Rekognition”, its notorious facial recognition technology.
Timnit Gebru, a leader of Google’s ethical AI team, explained why she thinks “facial recognition is too dangerous to be used right now for law enforcement purposes” – highlighting the disparity in error rates, especially that detection technology does not work for darker skin. Let that sink in for us in India, where not all of us are “Fair and Lovely” and face colourism everyday because of our dark skin.
For several years, activists have been crying hoarse about the abuse of these technologies. However, on March 11, the Indian government casually announced the full adoption of facial recognition technology enabled surveillance. We were told that using photographic and other information from government “databases”, 1,100 individual participants in the Delhi riots had been identified. The number was later raised to 1,900. The software, we are told “doesn’t see clothes”, only faces – an apparent attempt to deny that the technology used to identify individual “rioters” today could be used to identify any individual tomorrow.
Whether UIDAI’s Aadhaar database was among the government data stores to which this announcement referred has been the subject of conflicting statements. NatGrid hasn’t been specifically mentioned. It hardly matters. The government has actually announced its determination to use its data assets to put a face on every action to which the government objects.
As we’ve previously written, other advanced democracies have been slowing down or stopping altogether uses of facial recognition in the public sphere. They’re aware of the enormous harm that can be done by governmental abuse of the technology of automated facial recognition. From a moratorium in the EU to outright prohibitions by California municipalities, the current trend demonstrates a sudden, sharp recognition that a government able to identify every face in public everywhere all the time can extinguish freedom at a blow. This has been aided by the demonstration project in technological totalitarianism staged by the Chinese Communist Party in Xinjiang.
But here in India, we seem to be not only travelling at top speed in the other direction, the actions are indicating that rule of law is no more than a small bump in the road on our way. What such a use of facial recognition technology tells us is that government intends to flout both the provisions of the Aadhaar Act and the Supreme Court’s Puttaswamy judgment upholding our fundamental constitutional right of privacy.
It’s laudable for government to protect public order, to identify and prosecute criminals. A responsible government would seek present judicial authority for the novel use of privacy invading technologies, and would introduce legislation to formalise the processes of judicial review in future. A government which instead uses the occasion of public disorder to introduce such measures by executive fiat can only be suspected of taking advantage of events to subvert the Constitution it, and its police, are sworn to protect.
But now, the Covid-19 epidemic is being used as another “legitimate” justification for further tracking of public movements and interactions throughout society, with government taking another advantage of circumstances to extend its unaccountable use of this technology. Everyone’s justified fear of the spread of disease is being used by a government to travel faster and further down the road to totalitarian process, beneath the veneer of a democracy protecting itself in a time of crisis. The ineffective, botched job with the Aarogya Setu app speaks precisely to this.
Where government reads every face, political dissent is under permanent intimidation. We cannot live our lives outside the range of others’ cameras anymore. Wherever we go and whatever we do in the public sphere we can expect our image to be recorded without our knowledge or consent. Equally – as Aadhaar has escaped its original role in the protection of public subsidies against theft and fraud and become the universal link tying all chains of personal data together – we can expect our “data shadow” to follow us through every fiscal and administrative transaction. If government is allowed to put these two pieces together – by searching available images, in its own databases and social media collections and tying them to Aadhaar-linked databases of personal information – it has “total information awareness”.
Unless this awareness is subjected to judicial review when it is used, and there is subsequent judicial oversight to prevent abuse, our lives are no longer our own. The present government predicts – let us not say assumes – that the SC is too busy or too supine to impose the Constitution on the government’s will, rather than the other way around. Immediate intervention by the SC on the basis of petitions brought by advocates for constitutional privacy is necessary.
The disease is not our only fear. Unless we are vigilant now, even as we are put under so much additional strain by the pandemic and its human consequences, we shall live also through the death of freedom.
Mishi Choudhary is legal director of Software Freedom Law Centre, New York. Eben Moglen is Professor of Law and Legal History at Columbia Law School
DISCLAIMER : Views expressed above are the author’s own.