Popular generative artificial intelligence company OpenAI, best known for its ChatGPT tool, gathered some of the top tech policy and civil society representatives in India on February 12 to discuss AI-driven misinformation and election preparedness for the upcoming general election.
Former top Indian government Information Technology officials and tech scholars who attended OpenAI’s private roundtable meeting in Delhi said the Election Commission of India (ECI) could be doing much more in coordination with major tech platforms such as OpenAI to combat misinformation and disinformation during a time of heightened sensitivities.
The civil society briefing was attended by digital rights groups, law and policy experts, disinformation reporters, influencers, at least one former government official, and Big Tech firms’ representatives, multiple participants told The Hindu.
During the meeting, OpenAI officials emphasised that they were in the country to get a lay of the land and understand the most pressing issues surrounding AI, such as synthetic media (deepfakes) and misinformation. The outreach takes on added significance as India emerges as the firm’s second largest user base outside the United States.
Rakesh Maheshwari, a former senior director and group coordinator at the Ministry of Electronics and Information (MeitY), suggested to the company that it should engage in deeper consultations with the Indian government and the ECI.
“I know everyone in the room was focused on OpenAI products and tools and what can be used to improve things with misinformation but the real point is that the Election Commission of India should be doing more. We need to do more with them and bring them into this issue. That’s what our focus should be on right now,” Mr. Maheshwari, who has been director and group coordinator at MeitY from 2018 to 2023, said.
He said major online platforms and AI creators such as OpenAI, Google, Meta and others in conjunction with the elections commission need to agree on some rules around what is considered misinformation and how to communicate this clearly to each other. Misinformation notices should be received by online platforms or by the Election Commission and then the platforms have to act fast to stop dissemination of misinformation, Mr. Maheshwari said.
“Platforms like OpenAI, Meta and others can help enhance the Election Commission’s powers and point them to signs of misinformation and this can help reduce at least 30-40% of the problem which is enough for now,” said Mr. Maheshwari.
“This issue is heightened because of the national elections which are around the corner and so sensitivities are high and this problem should be accorded a high priority,” Mr. Maheshwari added.
Steps to safeguard misuse
While OpenAI was there on a listening trip, it told participants, it laid out some baseline steps that the company had taken to safeguard its products’ misuse during elections across the world: a prohibition on using its chatbot product to campaign for political candidates, a prohibition on discouraging participation in elections and on returning incorrect information on polling dates and locations.
Another attendee of the closed door OpenAI meeting suggested that political leaders should commit to abstain from deepfakes, pointing to an open letter by the Internet Freedom Foundation to the ECI, which made a similar suggestion.
“Some in the room mentioned correctly that there should be some self-policing self-imposed restrictions by politicians on DeepFakes and other AI driven misinformation. That may not be realistic, it may not be feasible, but it’s a good idea.” said Mr. Maheshwari said.
Tech giant OpenAI has expanded its outreach within India, where it currently has only one part-time employee based in the country. OpenAI is planning on organising AI summits in India to speak with software developers in multiple cities and expand its footprint in the country, chief strategy officer Jason Kwon said at an event over the weekend.
