'Scope for misuse of AI is enormous'

A 200-year-old old conversation between Michael Faraday and William Gladstone still holds water when it comes to the links between science and policy. Asked about the utility of the experiments that he was conducting, the British scientist responded to the politician,“Sir, there is every possibility that you would soon be able to tax it.”

The experiments that Faraday carried out led to the discovery of electricity, the most important driver of economic growth in the last two centuries. Possibly because of such history lessons, modern-day scientists now seek governmental interference in something like artificial intelligence that not only has huge commercial potential but also scope for much misuse.

“AI is something where eventually government regulations would be required. With the technology advancing in the current pace, there would be positives and negatives. When you aggregate large amount of data and learns from it, there are benefits to the society from medicine to transportation. But technology is moving too fast for the government to respond and society to comprehend. We should be making sure any innovation is done in a responsible way,” Subra Suresh, a professor at MIT and former president of Carnegie Mellon University told DH.

An alumni of IIT Madras, Suresh who was the first (and so far the only) Indian-American to become the director of the US National Science Foundation in 2010, is currently the president of the Nanyang Technological University, Singapore.

Asked whether government regulations could be counter-productive, Suresh said, “policies can have different flavours to it, from something as simple as taxation that addresses human behaviour, incentives and disincentives, to requirements in areas such as privacy and confidentiality AI is no different from any technology except that it affects every sector. It is important for the government to keep up with the technology.”

The scope for misuse is enormous. In a data-driven decision making process, feeding of wrong data would lead to wrong outcomes. Suresh said the government must play a role to not only ensure good quality data is flown to the AI systems, but also to convince people about sharing data without compromising on their rights to privacy.

“It's important to have data policies. By sharing data in a certain way while protecting individual rights and privacy, you can have collective benefits. It's a trade off,” he summed up.