Reuters
KEY POINTS
- Schumer acknowledged that AI is unlike any common topic that Congress has been addressing
- Senate Majority Leader wants forums to start as early as September
- The EU is approaching the final phase before its AI law becomes official law
Senate Majority Leader Chuck Schumer, DN.Y., has unveiled a strategy for lawmakers to regulate artificial intelligence (AI) as the United States seeks to catch up with the European Union in curbing the rapidly evolving technology.
Schumer conceded for the first time in a speech Wednesday that policymakers are “starting from scratch” because AI “is unlike anything Congress has dealt with before,” Semafor reported.
The veteran senator suggested that instead of holding hearings to work out possible regulations for the technology, Congress should set up “AI Insight Forums” that bring together technology leaders, AI critics, labor representatives and lawmakers.
Schumer proposed at least nine panels to identify and convene the most challenging AI regulation issues. According to CNN, the first forums should start as early as September, Schumer said.
The forums will reportedly cover various topics such as national security risks, AI workforce, copyright and intellectual property, and “protection from doomsday scenarios.”
Schumer also acknowledged criticism of the government’s seemingly slow pace in regulating the rapidly evolving technology. However, he pointed out that AI is a completely different topic than new issues that the Congress is facing. “Experts are not even sure what questions politicians should be asking,” he said, according to the newspaper.
The Democratic senator said he and his team held talks with more than 100 AI developers, academics, industry leaders and workforce experts to craft a legal framework.
For Schumer, there are five pillars that should be the focus of AI innovation. Innovation should come first, “but with safety, accountability, fundamentals, and explainability,” which Schumer calls “SAFE innovation for AI,” CNN reported.
He also mentioned that an agreement could be reached to limit state intervention in AI innovation to a certain level.
Schumer’s proposal came after Time reported on OpenAI’s lobbying efforts in the European Union to declassify general-purpose AI systems like its own ChatGPT chatbot as “high-risk” technology.
OpenAI has reportedly proposed changes to the bloc’s comprehensive AI law since last year, and some of the company’s proposed changes were later applied to the final text of the regulation, according to documents obtained by Time through freedom of information requests from the European Commission.
Sarah Chander, senior policy adviser at advocacy group European Digital Rights, said OpenAI “got what they asked for” and that it just shows how tech companies “mask” their financial interests in weakening regulations by using “the utility argument.” and public benefit” use AI,” according to Time.
While OpenAI is lobbying EU officials for softer AI rules, it is also urging governments to speed up work on regulating the technology.
Sam Altman, CEO of OpenAI, told lawmakers in May that the company believes “regulatory intervention by governments” is critical to ensuring the risks posed by AI are mitigated.
Former MEP Marietje Schaake dismissed the question of allowing tech leaders a strong say in the drafting of tech laws in an editorial for the Financial Times published earlier this month.
“Legislators need to remember that what business people care about most is profit, not social impact. Policymakers must not allow tech CEOs to shape and control the narrative, let alone the process,” she wrote.
Meanwhile, the United States lags behind the EU on AI regulation. Unlike the EU, whose AI law was passed by Parliament earlier this month, the US has yet to draft similar laws.
The White House under President Joe Biden has presented a “Draft AI Bill of Rights”, which, however, only serves as a “guideline for a society that protects all people” from the threats of unregulated AI.
The Biden administration has also launched investigations into possible accountability rules for AI systems and is asking the public for opinions on the matter.
The EU, on the other hand, only needs to get approval from all bloc member states before its AI law becomes official law. The EU government is aiming for an agreement with all member countries by the end of 2023.