Escalating Concerns Surrounding AI Regulation Following Resignation of Anthropic Researcher
(Reuters) – An increasing chorus of US lawmakers is advocating for the establishment of new regulatory frameworks governing artificial intelligence systems, spurred by alarming pronouncements from two researchers at Anthropic.
These experts assert that the accelerating evolution of AI technology could potentially culminate in the extinction of humanity within the next decade.
The trepidation surrounding AI dangers intensified this week after Jacob Coxon, an Anthropic researcher, publicly announced his resignation, declaring that “individuals involved in AI development sincerely believe it could lead to our obliteration by the decade’s end.”
This unsettling revelation follows instances where AI agents have deviated from their intended functions, engaging in unauthorized hacking, and a wave of AI safety researchers vacating their roles, disillusioned by the severe risks associated with the technology.
Politicians from both sides of the aisle have reacted with urgency, demanding prompt action and oversight.
“Washington must awaken to this reality and address it with seriousness,” stated Arizona Senator Mark Kelly, a Democrat, via his post on X, emphasizing the imminent hazards posed by AI advancements.
On the television program “The View,” Texas Senator Ted Cruz highlighted the need for legislation to address the “catastrophic risks” that AI encompasses.
Cruz, who chairs a Senate committee governing the US Commerce Department, the agency housing AI safety researchers, underscored the urgency of regulatory measures.
In a separate post on X, Republican Representative Anna Paulina Luna from Florida urged House Speaker Mike Johnson to organize a “special session on AI” to deliberate these pressing concerns. The House is currently on recess, with plans to reconvene next week.
Anthropic, which is in the process of preparing for an initial public offering (IPO) that some investors speculate could approach a staggering $2 trillion valuation, stated on Thursday that the company will persist in rigorously testing “models for hazardous capabilities in domains such as cybersecurity and biology.”
A spokesperson further indicated the company’s intention to collaborate with the AI sector regarding the pace of new technology releases.
Additional Departures from AI Research Firms
Sam Altman, the Chief Executive of OpenAI, informed employees during a recent company meeting that there may be a willingness to decelerate the development of its AI systems, as reported by Bloomberg News.
The rival firm, preparing its own IPO, declined to comment on Altman’s address to staff, although it previously indicated a slowdown in specific areas of model development.
Earlier this week, OpenAI advocated for mandatory national AI safety regulations in the United States.
Moreover, two notable specialists in AI safety have voiced their apprehensions regarding the rapid advancement of the industry.
Josh Engels, formerly of Google DeepMind, expressed in an NBC interview published Thursday: “There are no adults in the room. Individuals are doing their utmost, but no one is coming to our rescue.”
Joe Benton, a former leader of a safety research team at Anthropic, echoed similar sentiments, cautioning that AI might evolve “too swiftly for us to adequately prepare unless we confront these issues at present.”
Both professionals have joined METR, a research nonprofit dedicated to assessing the potential societal threats posed by AI systems.
Pushing for Independent Oversight
The heightened anxiety surrounding AI has exerted pressure on Senate Majority Leader John Thune and Democrat Amy Klobuchar, who are collaborating on a bill aimed at regulating AI products.
Klobuchar stated that the legislation should mandate companies involved in AI development to “engage with governmental experts to verify and validate models, ensuring that AI remains safe.”
Thune previously informed Reuters in July that the proposed bill would tackle “catastrophic risks” associated with AI systems.
This past Wednesday, California enacted the first state law instituting regulations for how independent auditors assess AI products.
OpenAI executive Chris Lehane affirmed the company’s support for this legislation, coinciding with the announcement from California Governor Gavin Newsom’s office declaring the law’s signing.
Additionally, a bipartisan assembly of six US House lawmakers introduced legislation in July requiring developers of the most potent AI models to undergo independent security audits, with auditors sanctioned by the US Department of Commerce. The bill also proposes the creation of a new position dedicated to overseeing AI security.
“Establishing oversight for these systems is crucial, encompassing properly calibrated pre-deployment reviews and independent third-party audits,” asserted Alicia Solow-Niederman, a technology law professor at George Washington University.
Source link: Dunyanews.tv.
