CEO of OpenAI, Samuel Altman, recently testified before the US Senate Subcommittee on Privacy, Technology, and the Law regarding the regulation of artificial intelligence (AI) platforms such as OpenAI’s ChatGPT. During the hearing, Altman emphasized the need for the United States to develop regulations that balance AI safety with ensuring access to the technology’s numerous benefits.
The session commenced with the playback of AI-generated audio featuring Senator Richard Blumenthal (D-CT). Blumenthal disclosed that the audio was composed using AI voice cloning software, which utilized his previous appearances in Congress. The script for the audio was generated by ChatGPT based on its analysis of Blumenthal’s congressional record. This demonstration aimed to highlight the capabilities of AI technology while also emphasizing the potential risks and harms associated with it. Blumenthal cautioned that AI voice cloning could be exploited to create false audio, such as fabricating statements endorsing foreign leaders or compromising national security. He stressed the urgency for the United States to address this reality.
Senator Josh Hawley (R-MO) echoed Blumenthal’s concerns, emphasizing the rapid advancement of AI technologies like ChatGPT. He noted that a hearing of this nature would not have been feasible a year ago due to the limited existence of such technology.
To assist lawmakers in navigating this rapidly evolving field, a panel of academics and experts, including Samuel Altman, IBM Chief Privacy and Trust Officer Christina Montgomery, and New York University Professor Gary Marcus, provided testimony. The panel collectively emphasized the United States’ need to establish laws and regulations governing AI. They proposed the implementation of licensing and testing requirements to prevent AI technologies from promoting harmful content or misuse. Montgomery and Marcus went as far as suggesting a six-month halt on AI development to enable governments to develop a comprehensive regulatory framework.
Find the legal job that fits your lifestyle and career goals with BCG Attorney Search.
Currently, the AI industry operates primarily on voluntary adoption of rules and regulations by individual companies. Altman mentioned that OpenAI has established guidelines for the responsible use of ChatGPT. However, lawmakers expressed apprehension towards relying solely on voluntary measures, citing previous instances of social media companies, such as Meta and TikTok, failing to adopt user protections without congressional oversight. These companies’ lack of regulation led to increased scrutiny and concerns over user safety. Blumenthal stressed the need for Congress to act on AI regulation to prevent a similar “race to the bottom” scenario.
One critical concern highlighted by the panel was AI’s capacity to manipulate and persuade individuals through disinformation. Altman and Hawley specifically emphasized the importance of preventing AI from disseminating disinformation ahead of the 2024 US presidential election. The potential for such harm had already prompted Italy to ban ChatGPT in March.
Altman acknowledged that while there may eventually be a point where the public can readily identify AI-generated audio, images, and text, similar to recognizing photoshopped images, that point has not yet been reached. At present, Altman described AI as “photoshop on steroids,” highlighting the significant impact and potential risks associated with AI-generated content.
The testimony before the US Senate Subcommittee shed light on the pressing need for AI regulation. It highlighted the potential harms posed by AI technologies, particularly in terms of disinformation and manipulation. The panel of experts emphasized the importance of developing comprehensive laws and regulations to ensure the responsible and ethical use of AI. The discussion underscored the necessity for governmental involvement in establishing a robust regulatory framework that safeguards against the risks associated with AI while harnessing its transformative potential.
