As an international grouping took a first step Wednesday in setting up a common global framework to regulate artificial intelligence (AI), experts cautioned that such agreements would only work between like-minded countries.
And that is the story of the Global Partnership on Artificial Intelligence (GPAI) — an attempt largely led by the democratic world to adopt a global risk-based approach to AI.
It includes countries like India, the US, the UK, France, Japan and Canada, among others. But China, a tech superpower, is not part of the multilateral grouping — and, a senior government official said, is unlikely to be invited into the fold.
ADVERTISEMENT
On Wednesday, the GPAI Summit adopted the New Delhi declaration, underscoring the need to mitigate AI risks and promote equitable access to critical resources.
Rama Vedashree, advisor at Dallas Venture Capital, said that while regulatory synergy will naturally emerge globally, it will only happen between “rights-respecting countries”.
“Some principles will emerge and there would be a synergy between various regulations — at least among all the rights respecting countries. I would not say (there would be regulatory synergy in) all the countries that are ahead in terms of technology. Whether you look at the declaration at the UK AI Safety Summit or even the G20 declaration, which had adequate focus on AI and other technologies, there would be a synergy that will come,” Vedashree told The Indian Express.
Thomas Schneider, chair of the Council of Europe Committee on AI, said while it may be easy to solve the smaller challenges related to AI, the bigger ones, where a number of stakeholders are involved, may be much harder to address. He pointed out that global AI regulations can be interoperable only among countries that share the same values.
ADVERTISEMENT
“We will solve the small problems easily. But the big challenges, where many people are involved or everybody’s involved, are much harder to solve… you have to share more or less the same values in order to have interoperable implementations of the values,” he said.
Schneider admitted that the proliferation of AI systems can not be contained, but batted for a framework that can prevent the weaponisation of AI-based systems during wars. “We will have to use these technologies because of competition issues, or simply because people are curious to use them… hopefully we also find a framework that ensures that these technologies are not used to kill each other in war,” he said.
Asked if regulators should prioritise innovation or safety related to AI systems Schneider said: “It should not be an either, or.” He emphasised that regulations need to offer the right incentives to various stakeholders.
ADVERTISEMENT
The EU has introduced its own AI Act. Asked for his recommendations on what India and other countries should prioritise in their own AI-related laws, Schneider said: “…we should respect fundamental values, human rights, not fight against democracy, against the autonomy of the people, against peace”.
Sandip Patel, Managing Director at IBM India & South Asia, said his company welcomes “precision regulation” on AI that would not limit people from using such systems. He also said declarations such as the one adopted at the GPAI Summit can serve as a blueprint for countries to frame regulations as per their specific market needs.
“With declarations like these, like the one that was signed in the UK as well, I think there’s a significant opportunity for countries that are advanced in these technologies and have taken a point of view in terms of some of these policies and frameworks and standards, to create some form of a global framework… that can be used with specific country specific, market specific regulation especially for countries that are ahead in this space,” he said.
“We as a company believe in precision regulation… I do believe that at least for technologies like AI, there is a lot of room to drive precision regulation as to what really needs to be regulated as opposed to limiting anyone from using AI,” Patel said.