What India Should Remember When it Comes to Experimenting With AI

India is rapidly digitising. There are good things and bad, speed-bumps on the way and caveats to be mindful of. The weekly column Terminal focuses on all that is connected and is not – on digital issues, policy, ideas and themes dominating the conversation in India and the world.

Global information technology markets have been marked with a slowdown over the past year, resulting in uncertainty with the future of this industry and the jobs within the sector. This slowdown in growth is forcing capital markets to look for alternative growth opportunities and they seemed to have found their answer in artificial intelligence.

The World Economic Forum in January this year was abuzz with ChatGPT and how AI is here to transform work and industries. The big question everyone seems to be having is how will AI affect society. 

With a huge infusion of capital into companies building AI, every company out there is hunting to find one problem they can solve to demo the potential of AI. There is going to be an explosion of companies that want to push AI in every aspect of the economy and governance. The early days of AI development and deployment is going to be full of trial and errors just like it is with any software development process. 

It has been a policy of the government of India to promote IT businesses in the country by allowing companies to experiment with new technologies in governance, so that they can export them after their development.

The push for e-governance in India was primarily to help grow a domestic IT industry that is capable of working in real world environments. While this benefits the IT industry, the experimentation leads to dehumanising effects on the population. 

Be it any new emerging technologies like blockchain, facial recognition and AI, state policies have been so far oriented towards helping IT industries to maintain India’s economic growth. But these policies have cost the people who were being experimented on time and again. From Aadhaar to CoWIN, the early experimentation of technologies in India to help export them has had its pitfalls on the population. This should not continue with the new wave of AI push. 

Also read: Kashmir: CoWIN Sends ‘Fully Vaccinated’ Texts to Several Who’ve Taken Just One Shot

The private sector is more than welcome to develop and test new technologies in controlled environments. But the push of experimental emerging technologies on the live population – without the right testing and safety checks – needs to be stopped. There has been a push by the private sector itself to create regulatory sandbox environments where regulators can monitor experimentation of new technologies. It is more than welcome for the government to create these controlled environments for experimentation, but any large-scale implementation needs sufficient checks and balances. 

It is not techno-solutionism that is at the heart of the problem of why we see technologies fail when being experimented on India’s poorest. It is state apathy towards its population that is continuing these broken experiments. Even after continuously pointing out errors and design flaws with some of the largest technology infrastructures, there is a lack of introspection. There is no room to negotiate with the government or with the companies that produce these technologies. 

Technology has dehumanising effects where humans do not have any mechanisms to appeal to what they are being subjected to by the machine. With AI, these effects of de-humanisation are only going to increase. But to confuse these as effects of choices made by an artificial created entity is fallacy. These are social choices that are being determined by social actors. The choice to force facial recognition, AI-based nutrition tracking of school children, and to push biometrics on the population are not technological effects or choices. 

Also read: Making Aadhaar-Based Payments Compulsory for NREGA Wages Is a Recipe for Disaster

The Indian IT industry and government is deeply concerned with AI replacing jobs in our IT service sector and it threatens them. It is surprising that of all people, the tech evangelist who championed Aadhaar and IndiaStack, Sharad Sharma, is calling on people to be cautious of AI and how it is deployed in India. I appreciate Sharma’s words, but they are not words of selfless interest. In all class societies, production of technologies is not isolated with the interests of just one industrial class. The Indian IT industry’s tryst with AI will affect not just this one class, but our entire society. 

We cannot afford to have a ‘One Architect, One Solution’ approach towards development of AI, where the IT industry decides how it should be developed or deployed. The lack of social participation will be a serious problem with AI, if not other technologies. Any AI that is developed without checks will lead to social biases being pushed as AI biases.

The only way to have a larger consensus in production of AI is to get people to be part of it, rather than risk it being delayed, broken and under-performing compared to Silicon Valley. While there is a need for national bodies to help develop AI and standards around them, it cannot be led only by the private sector or be allowed to be taken over by them. 

Srinivas Kodali is a researcher on digitisation and hacktivist.