Putting AI in its place

AI technologies are made, sold, and used by humans; therefore, the scope for responsible action clearly exists at every step.

Awe and fear dart through public discussions of AI. Its fast-enveloping presence across the spheres of human activity stokes the ‘What next?’ in the mind. Overjaunty pronouncements about the imminent loss of millions of jobs due to generative AI have gone on to make the future of employment appear shaky. Meanwhile, the ubiquitous cliches about how every technology has pluses and minuses ask everyone to sit back and get used to AI. All add up to make AI seem a natural, unstoppable force.

AI technologies are made and sold and used by humans: the scope for responsible action clearly exists at every step. Therefore, what these technologies seek to accomplish, how they become available to users, and their consequences for the democratic life of societies invite public discussion. Just when colossal profits and international power games appeared to drive AI innovations without any regard for ethics, the recent refusal of Anthropic, the US-based AI developer, to allow the US Defence Department to deploy its AI technology for making autonomous weapons that decide independently of human supervision and for citizen surveillance has brought the focus on the ethics of AI.

The cultural attitude towards AI needs to become less fearful and quiescent, and more demanding of social accountability. The owners and creators of AI technologies do strive to make them increasingly reliable. But this pursuit of reliability must be accompanied by efforts to regulate them effectively, implementing policies that prevent socially destructive consequences.

Consider the prospect of mass job losses, which hangs over public discussions of AI like a ghost. Scholars of this subject note that widespread layoffs are unlikely in the near future, as the adoption of AI takes time. Moreover, the disappearance of some jobs may be accompanied by a rise in demand for new jobs in emerging areas. Well-organised employer unions can also negotiate with management to shape how AI is implemented in the workplace, ensuring that human workers remain meaningfully engaged. And there will be tasks where AI simply cannot take the place of humans.

Just as pensions, welfare taxes, the regulation of working hours, industrial safety standards and other kinds of social protection emerged in the 19th century in response to the radically new circumstances brought about by industrialisation, new kinds of social insurance will become necessary to imagine to address the wide-ranging social transformations ensuing in the wake of generative AI. What new responsibilities can the governments take on to address the AI-induced transformations in the labour market and in society at large? How can businesses making gains from AI play a part in supporting the new social insurance measures that will become necessary? What new social, economic and political rights can the citizens envision to secure for themselves dignity and well-being in the changed times? Questions like these need to engage the public imagination alongside all else.

The energy-guzzling AI data centres have added a new energy dynamic in the present, making the sustainable use of AI a paramount concern.

The impact of AI, as well as the challenges of regulating it responsibly in countries like India, where digital technologies are not accessible to all and where a large number of workers operate outside the formal economy, are distinct and require policy measures grounded in these realities. Some of the policy shibboleths will need to be rethought. For instance, the pathetic urge among the planners to shift the “surplus-population” from rural areas to cities by training rural youth for service occupations in cities will stand out as unwise. What’s the point of training them for jobs that AI will dispense with? Measures to restore and create livelihood options in rural and tribal areas make better policy sense in the age of AI.

What if AI can think on its own? Or interact with each other on their own? Geoffrey Hinton, the Nobel laureate and ‘Godfather of AI,’ recently urged that the AI safety institutes that make AI benevolent and desist from eliminating human beings be built, a goal that he feels can be shared by rival countries even as they compete to make their AI smarter than those of the others. Let’s keep thinking.

(The writer is a Vidyashilp Professor who looks for new ways of looking)

Disclaimer: The views expressed above are the author's own. They do not necessarily reflect the views of DH.