Five ways AI can be regulated
Although their efforts to keep up with developments in artificial intelligence have largely failed, regulators around the world are taking starkly different approaches to overseeing the technology. The result is a highly fragmented and confusing global regulatory landscape for a borderless technology that promises to transform the labor market, contribute to the spread of disinformation, or even pose a risk to humanity.
The main frameworks for regulating AI include:
The European risk-based law: The European Union’s AI law, which is being negotiated on Wednesday, sets rules proportionate to the level of risk posed by an AI tool. The idea is to create a sliding scale of regulations aimed at imposing the toughest restrictions on the riskiest AI systems. The law would categorize AI tools based on four designations: unacceptable, high, limited and minimal risk.
Unacceptable risks include AI systems that perform social scoring of individuals or real-time facial recognition in public places. They would be banned. Other tools that pose less risk, such as software that generates manipulated videos and “deepfake” images, should reveal that people are seeing AI-generated content. Violators can be fined 6 percent of their worldwide turnover. Minimally risky systems include spam filters and AI-generated video games.
US Voluntary Codes of Conduct: The Biden administration has given companies the space to voluntarily monitor themselves for safety and security risks. In July, the White House announced that several AI makers, including Amazon, Anthropic, Google, Inflection, Meta, Microsoft and OpenAI, had agreed to self-regulate their systems.
The voluntary commitments include third-party testing of tools known as red-teaming, research into bias and privacy concerns, sharing risk information with governments and other organizations, and developing tools to address societal challenges such as climate change while also including transparency measures to identify AI-generated material. The companies have already fulfilled many of these obligations.
US technology-based law: Any substantive regulation of AI will have to come from Congress. Senate Majority Leader Chuck Schumer, Democrat of New York, has promised a comprehensive AI bill, possibly next year.
But so far, lawmakers have introduced bills aimed at the production and deployment of AI systems. The proposals include creating an agency like the Food and Drug Administration that could set regulations for AI providers, approve licenses for new systems and set standards. OpenAI CEO Sam Altman has backed the idea. However, Google has proposed that the National Institute of Standards and Technology, founded more than a century ago without regulatory powers, would serve as a center of government oversight.
Other bills target copyright violations by AI systems that gobble up intellectual property to create their systems. Proposals have also been made on election security and limiting the use of ‘deep fakes’.
China is moving fast in regulating speech: Since 2021 China has made rapid progress in rolling out regulations on recommendation algorithms, synthetic content such as deep fakes and generative AI. For example, the rules prohibit price discrimination by recommendation algorithms on social media. AI creators must label synthetic AI-generated content. And draft rules for generative AI, like OpenAI’s chatbot, require that training data and the content the technology creates “true and accurate”, which many see as an attempt to censor what the systems say.
Global cooperation: Many experts have said that effective AI regulation will require global cooperation. So far, such diplomatic efforts have produced few concrete results. One idea that has been put forward is the creation of an international agency, similar to the International Atomic Energy Agency, created to limit the spread of nuclear weapons. One challenge will be overcoming the geopolitical mistrust, economic competition, and nationalistic impulses that have become so intertwined with AI development.