Future of AI Regulation – Updates on AI & Technology by Outreinfo
OpenAI CEO Sam Altman testified before Congress on May 16, 2023, calling for regulators to set limits on powerful AI systems. Altman acknowledged the potential for significant harm and agreed that government oversight is critical. Governments worldwide are now debating regulation of AI in future. The question for business leaders is not how or when AI will be regulated, but by whom, as this will determine the speed and trajectory of AI’s transformation of the global economy.
Since the release of OpenAI’s ChatGPT in November 2022, the use of generative AI has exploded. ChatGPT reached one million users in five days, surpassing the growth of Facebook, Spotify, and Netflix. LLMs like Mi journey and DALL-E have also become popular, generating millions of images daily. Generative AI is a “Big Bang Disruptor,” offering a better and cheaper experience from the moment of release. The potential for LLMs is vast, potentially revolutionizing multiple industries as an uber-disruptor.
Given the potential scale of disruption from AI, as well as concerns about privacy, bias, and national security, it is reasonable for lawmakers to take notice. Many, including developers like Altman, are looking to lawmakers to fill the role of the sorcerer in Goethe’s poem “The Sorcerer’s Apprentice,” where the sorcerer restores order after his apprentice unleashes forces that threaten to destroy everything.
In the United States, multiple actors are competing to lead the regulation of AI in future. Senate Majority Leader Chuck Schumer is calling for pre-emptive legislation to establish regulatory “guardrails” on AI products and services, focusing on user transparency, government reporting, and aligning with American values. The Biden Administration is also working to implement a White House blueprint for an AI Bill of Rights, calling for safe and effective systems that don’t discriminate or violate privacy and offer human fallbacks.
The National Telecommunications and Information Administration (NTIA) at the Department of Commerce has opened an inquiry about the usefulness of audits and certifications for AI systems. The agency is seeking comments on accountability for AI systems, including assessment, certification, and auditing.
Federal Trade Commission Chair Lina Kahn claims her agency already has jurisdiction over LLMs and warns that AI could exacerbate existing problems in tech, including collusion, monopolization, and unfair competition. Kahn also believes generative AI risks turbocharging fraud and could violate privacy and anti-discrimination laws.
AI-related legislation has been introduced in at least 17 states, with some proposals incentivizing local development of AI products and others limiting its use in applications such as healthcare and hiring. Many states are considering creating their own task forces to recommend future legislation.
However, there are few specifics in these proposals, and regulators are likely to have little impact on the technology’s development in the short term. The application of new rules will be a matter for the courts, which could take years. Governments have also had difficulty attracting the technical expertise required to define the potential harms of AI.
The Department of Commerce is credited for asking the right questions about AI regulation, but it is unclear if Secretary Gina Raimondo has the legal authority or political clout to create a sustainable certification process or gain support from the tech industry. The Department’s inquiry is part of a larger White House effort to create a trusted environment for AI services, which would require unprecedented coordination and cooperation across government silos.
Discussions about AI regulation are taking place against the backdrop of significant changes in American law that will determine who will ultimately regulate AI in future. Recent Supreme Court decisions have shifted power from federal regulators to the courts and states, adding fragmentation, uncertainty, and delay to enforcement actions. The Court has allowed businesses to challenge agency rulemaking by demanding more specific instructions from Congress, effectively outsourcing the final decision on adopted rules to federal judges. Meanwhile, the technology continues to evolve rapidly.
Given the limitations of the US regulatory system, major AI regulation is more likely to come first from outside the US. In competition law, momentum has already shifted from the US to Europe, where the EU has passed substantial new internet legislation while Congress has been slow to act. The European Parliament recently approved the AI Act, which would ban applications with “unacceptable” levels of risk, require pre-approval and licenses for others, and impose substantial fines for violations.
Regulators in China are also moving quickly to incentivize home-grown AI products and services and define their operation, potentially limiting non-Chinese companies’ interaction with Chinese users and setting a de facto legal regime for future applications.
It is uncertain if government action can effectively balance the value of AI in future with its potential harm to the economy or society. The ability of governments to regulate LLMs will likely fall short due to the incremental nature of law and the exponential evolution of technology. In the interim, business leaders and academics should follow the Department of Commerce’s initiative and develop nongovernmental regulators, audits, and certification processes to incentivize ethical and trusted AI products and services and identify reliable applications.
There is a long history of successful self-regulatory bodies, such as medieval merchant courts and modern groups like the International Standards Organization, which develop and certify compliance with standards and best practices. Similar efforts have addressed corporate standards for dealing with authoritarian regimes and the development of internet software and protocols. While some government regulation is inevitable, the most promising way to avoid provoking the sorcerer is to avoid making too big a mess in the first place.
