Somewhere between the hype and fear of artificial intelligence, AI governance exists or will exist. Artificial intelligence-powered chatbots, image generators, and search engines have become ubiquitous, with dire warnings issued by many of the technology’s developers about the unrestricted development of the technology. Calls for policy intervention have increased in recent weeks, with regulators looking to step up their efforts.
Hundreds of engineers and researchers have issued multiple open letters warning of the dangers of AI, with a letter published in late March advocating a six-month “pause” for the development of new AI models. doing. Most recently, veteran scientist Jeffrey Hinton, often referred to as the godfather of AI, resigned from his role at Google with an equally dire prognosis.
“I don’t think they should expand this further until they understand if they can control it,” Hinton said. new york timesAs AI systems become more sophisticated, he (and others) say they could disrupt the global job market, distort online reality, and, at worst, surpass human intelligence. We are calling for global regulation and cooperation to curb the technologies that one of the It’s possible, he says.
Setting a universal regulatory framework for technology has always been difficult, but it is becoming more difficult as technology advances. Negotiations on all rules, from social media to 5G cellular technology, are riddled with geopolitical issues and disagreements over the best approach. AI has great potential to transform economies and societies, not necessarily for the better, but it presents unprecedented challenges.
As is often the case with new technology, Europe is at the forefront of regulation. European Union lawmakers are set to vote next week on the bloc’s AI law, first proposed two years ago and likely to come into force two years later. The law designates “high-risk” applications for artificial intelligence, such as law enforcement, critical infrastructure, education and employment, and imposes stricter compliance and testing requirements on companies creating and deploying these applications. Subject to requirements.
One problem is that technology moves faster than regulation. Over the last few years, artificial intelligence has shown that it can mimic human intelligence using so-called large-scale language models. enjoy a salad.
“When the AI Act was written, there was no generative AI or large language models,” said Gerard de Graaf, the EU’s Senior Digital Envoy to the United States. European lawmakers are now scrutinizing the proposed bill. “We are not going to negotiate another deal, so this will have to stand the test of time,” he said.
Across the pond, US AI regulation is still a work in progress. Last year, the White House released an “AI Bill of Rights Blueprint” outlining his five principles to prevent discrimination and protect user privacy and safety, and the National Institute of Standards and Technology released his January Released his AI risk management framework. Congress has also begun to mobilize, with Democratic majority Senator Chuck Schumer launching an effort last month to craft a comprehensive AI law.
But so far Washington has taken a voluntary approach to compliance, with experts saying a more binding approach to AI regulation is needed.
“What we have learned from the crises and impacts of the last decade is that soft regulation is not enough to regulate this sector. Sarah Myers-West, former senior adviser to
Some regulations, such as copyright, privacy, discrimination and data protection laws, have already been enacted and are still in force. Temporarily Banned ChatGPT is a very popular chatbot developed by Microsoft-backed company OpenAI.
Washington and Brussels are laying the groundwork for global AI governance through a bilateral Trade and Technology Council. Digital ministers from the G-7 countries also dedicated a large part of last weekend’s meeting to “Responsible AI and Global AI Governance”, endorsing a risk-based approach similar to EU law. The United States also stepped up efforts to develop a global framework, releasing a National Standards Strategy for Critical and Emerging Technologies on Thursday, including a section on the topic of artificial intelligence.
“We are inviting more international partners and allies to support our common standard framework,” a senior government official told reporters.
The big challenge, of course, is the dragon in the room. The control of China’s technology sector contrasts with the open, global Internet, where there is no “Great Firewall” in the West. And with AI, Beijing’s behavior doesn’t seem to change. Like social media and big tech companies in general, China has far stricter requirements on how AI companies collect data, train algorithms, and produce output consistent with Beijing’s censorship and government controls. has established regulations to impose
It only took China days to restrict ChatGPT when the program began taking the world by storm earlier this year. , video, images, text, etc. Matt Sheehan, Fellow of the Carnegie Endowment for International Peace, whose research has focused on China’s technology sector According to the company, this is a way to iteratively respond to and continue to respond to technological changes while increasing oversight.
“All previous regulations, looking back at Chinese policy documents and Chinese media, are clearly rooted in fear of losing control over the flow of information,” Sheehan said. “We are very broadly supportive of the relationship between the Chinese government and AI, but we would like to cover the rationale for information management.”
China has long-standing ambitions to become a world leader in AI. It could offer an alternative vision of technology development and use that more authoritarian (or less democratic) states could choose to follow, rather than the EU’s more risk-based approach to regulation. I have.
“The situation in the countries where the regulations are deployed is very important,” Myers-West said. “So to that extent [EU’s] The AI law will be a template for other countries to adopt. It is also important to look at the constitutional protections, rights and democratic conditions of the country adopting the law. Notably, this risk-based approach does not require neat migration to other types of jurisdictions. ”
According to Sheehan, the conversation about AI regulation with China may reflect what happened with climate change, with Beijing initially frustrated with the rules imposed by the West, and eventually with climate change. They tried to curb their own progress before deciding that was the problem. Bigger than the great power race.
“If that awareness continues to grow in both countries, we may be able to move towards such a shared understanding, or push towards non-proliferation guardrails,” he said. “At the moment, the next year or two for him seems like a hell of a lot, but beyond that, who knows.”
This does not mean that officials want China out of the conversation.
“We are not excluding any country when discussing the development of this standardization strategy. said Wednesday. “I think it would be a really bad outcome, in any scenario, for the world to divide into standards developed in different regions that don’t serve the American economy.”
But it has already happened. The technology is still in its early stages, but a patchwork of different regulations with broad priorities is emerging around the world, with wide gulfs between East and West, according to a report released last week by the Brookings Institution. 30 countries including USA, Singapore, China, Russia, Mexico and India. “The East is largely dedicated to building research and development capabilities, largely ignoring the traditional ‘guardrails’ of technology management,” the authors write. “By contrast, the West is dedicated to ensuring these guardrails are in place.”
One of the other risks is entangling AI development into the wider US-China rivalry and potential economic and technological decoupling. Some suggest that China’s efforts to develop AI should not hold back the United States over concerns about how the technology could be misused. As Myers West suggested, that would be a mistake.
“There is a growing amount of arms race rhetoric being employed around the development of artificial intelligence, which seems to justify an arms race with China as a reason not to proceed with strengthening or enforcing existing regulations.” she said. “So I think there are real reasons to proceed cautiously about this rhetoric and consider who benefits from it.”
On the flip side, there are real concerns given the rapid pace of AI development and the potential for AI to become centrally important to the knowledge industry in the next few years.
“Regulators are always behind market developments,” said de Graaf. “The problem is when regulation lags so far behind the market that it becomes an obstacle to progress and innovation.”
Yet even Edmund Hoyle’s rules have been revised and updated over the years.
“This is not an incremental change, nor does it affect one sector. This is a transformation. “So policy makers in all sectors are going to look at the rulebook and ask themselves if it is still fit for purpose. That is what AI is doing for all of us. ”
