Raj N. Chandra Shekhar
In the space of seven days, two news stories emerged that encapsulated the debate over artificial intelligence (AI) regulation. The first was a little-known story about OpenAI eliminating its non-voting observer seat on its board of directors. The second was European Union unveils AI billIn their official journal.
While a direct correlation between the two would require a propaganda push that we are far from capable of delivering, the fact that Big Tech and AI regulation are unlikely to go hand in hand has not gone unnoticed by observers. Since last year, big companies have been sounding the alarm, arguing that AI legislation would undermine the EU’s competitiveness in the field and lead to capital flight.
Ironically, the US actually countered that the EU law would only benefit big tech companies with access to vast resources to cover compliance costs: the US pointed out that its own law focuses on risks around how AI models are used, whereas the EU regulation is about how these are developed.
Innovation that puts ethics first
Contributor to Harvard Business Review A few months ago, Reid Blackman, an AI author and ethics advisor, said the AI Act is a milestone on the road to responsible AI. The Act is not perfect, and will draw compliance complaints from all kinds of companies. “But it's a powerful regulation, and it's part of a growing movement to protect society from the worst of emerging technologies like AI,” he said.
Blackman noted that companies must continue to prioritize innovation, but also make ensuring regulatory compliance a priority with all senior AI leaders to protect their organizations from a loss of trust due to ethical violations. It doesn't matter if it's in the EU or elsewhere. (emphasis ours).
So what are Microsoft and OpenAI doing?
The last phrase above is an important one that requires global attention. And this is why I feel less optimistic about Microsoft taking an observer seat on the OpenAI board after a board coup, then abandoning it when the coup was overturned. Readers may remember that the OpenAI board fired Sam Altman, only to rehire him after Big Brother Microsoft stepped in.
Of course, the $13 billion investment would have helped Microsoft redraw the boundaries in its favor, firing top researchers Andrei Karpathy and Ilya Sutskever and installing people known to be quite opposed to regulation. It didn't go unnoticed by us, however, that Sutskever founded a new company focused on AI safety, Safe Superintelligence.
However, it's worth mentioning that despite abandoning its observer status, Microsoft still owns 49% of OpenAI. This is a sure way to incur the wrath of EU antitrust regulators and may also fall within the scope of the new AI law if enacted. So is this the start of a new anti-regulatory battle or a slick move towards compliance? Only time will tell.
OpenAI, on the other hand, didn't want to keep us guessing: They confirmed that following Microsoft's departure, there will be no observer on the board, denying rumors that Apple would claim the position. “We appreciate Microsoft's expression of confidence in our board and the direction of the company,” OpenAI said in a statement. Quote from TechCrunch.
EU AI law is a milestone, not a destination
Therefore Our Report As for whether the EU law could become the basis for similar regulations in other parts of the world, the United States, which aims to play its usual big brother role in international affairs, seems to be lagging behind on this issue, but there is a sample in the United States in the form of a bill in California aimed at regulating large-scale cutting-edge AI models.
California Senator Scott Weiner said the bill would require developers of large, powerful AI systems to adhere to “common-sense” safety standards, but opponents say the legislation would surely kill innovation and spell the end for the AI industry. The bill seeks to hold big tech companies responsible for potentially catastrophic dangers, something Meta, Gemini, OpenAI and Anthropic apparently don't want to accept in their charters.
Regulate now, not after something has failed
This is where the words of Helen Toner, director of strategic and basic research grants at the Center for Security and Emerging Technologies, come into play. AI players don't have to worry about regulations for now, but they should be more afraid of regulations that may suddenly come into force later.
Toner, the former OpenAI executive, said he understands the fears around regulation, but delays at the government level now could result in regulations being enacted once the industry is well established. Take the examples of social media and cryptocurrency: The uproar and immediate objectives could lead Congress to pass excessive regulation.
In fact, Tonner says startups and venture capitalists should set aside their fears about regulation and actively participate in the regulatory debate: Companies of all sizes should consult with lawmakers and lobbyists to ensure that any regulations that emerge are shaped by discussions with all stakeholders, not knee-jerk reactions.
While the US Congress has already held hearings on possible AI regulation, Tonner isn't sure if we'll see regulation in the country anytime soon, regardless of how the election cycle plays out. That's why we should expect lawmakers and industry in India to come together to discuss responsible AI for the future, not just in how AI models are created but also how they are used.
