Big Tech platforms like Meta and Microsoft are casting themselves as responsible users of Artificial Intelligence technology, while simultaneously seeking — and seeking to avoid stringent forms of government regulation.
Global Big Tech firms' efforts to keep regulation away from their Artificial Intelligence (AI) initiatives have succeeded in warding off any immediate action in India. Indian executives for Facebook parent Meta, Microsoft, and IBM on Saturday articulated the industry's overall point of view on regulating AI in a panel discussion. While each executive welcomed some regulation, the industry overall appeared eager to set the terms of discourse leading to rules that would eventually apply to them.
"At this stage, if we try to regulate AI very much, it will hamper the development of the sector also," Akash Tripathi, the in-charge for the Ministry of Electronics and Information Technology's Artificial Intelligence & Emerging Technology group said during a panel discussion. "We really need to create an enabling environment which is not focusing too much on regulation … It's a critical balancing act."
In May, Microsoft published a "blueprint" for regulators seeking to impose curbs on AI development. "You have to make sure that privacy and security is in-built" into AI models, Ashutosh Chadha, the Country Head and Director of Global Government Affairs for India at Microsoft said. "Those issues are fundamental, so our document [says] we shouldn't allow this tech to remain unwritten; there needs to be a government mechanism. It cannot be left to the private sector."
IBM, for its part, cast itself as a responsible player in AI. "One must look at [AI regulation] with a lot of optimism," Kishore Balaji, the Executive Director for Government & Regulatory Affairs at IBM in India said. "Having grandfathered technologies like this by way of executing AI models in critical and sensitive sectors like banking and finance, we have a responsible AI approach," he added. Mr. Balaji called on regulators to draw a distinction between developers of AI software, deployers, and users, arguing different approaches needed to apply for all of them.
Sunil Abraham, Public Policy Director at Meta India for Data Economy and Emerging Technology, did not speak on what government regulations should look like, but spoke of Meta's open sourcing of its AI models, and its refusal to do so for sensitive software like Voicebox, a model that can artificially synthesise a human's speech and make them say anything with just a few seconds as a sample.
Jason Oxman, the president of the Washington, D.C. based Information Technology Industry Council (ITI), was far more forthright. "Those who are calling for more regulation or expressing concern about AI are those industries disrupted by this," Mr. Oxman asserted. "That's not unusual."
"There was a guy named Luddite in the UK who led a group of agricultural workers against the advent of mechanized cotton picking," Mr. Oxman said (the movement was supposedly inspired by a figure named Ned Ludd, but his historicity isn't firmly established). "The Luddite burned down the new machines, worried about the loss of jobs. We don't want Luddites burning down AI."
Mr. Balaji of IBM said that there should be consistency across jurisdictions. "We're at a pivotal moment, and at a great place to look at this from a digital trade lens, and how there can be greater homogeneity on regulating," he said.
Regulations specifically on AI will likely be introduced as a part of the Digital India Bill, the IT Ministry's long-discussed legislation replacing the Information Technology Act, 2000. Minister of State for Electronics and IT Rajeev Chandrasekhar has indicated multiple priorities for the Bill, including a limited "harms-based" approach to regulating AI. However, no initial draft of the Bill is yet present.