Dario Amodei, CEO of Anthropic, converses with Linsey Davis of ABC News.ABC News
The advancement of AI has been remarkably rapid, prompting Dario Amodei, the chief executive of Anthropic—a prominent company in the field of artificial intelligence—to advocate for enhanced oversight of this technology, including governmental involvement.
In a candid discussion with ABC News, Amodei emphasized the necessity of developing AI with appropriate safeguards to ensure it benefits society.
“We are advocating for more robust regulation of the technology, suggesting that the government be empowered, in a specific manner, to prevent the deployment of hazardous AI,” he stated.
Dario Amodei, CEO of Anthropic, converses with Linsey Davis of ABC News.ABC News
“We must determine how to create this technology responsibly,” Amodei remarked. “However, the primary responsibility lies with us.”
Amodei’s remarks come in the wake of a letter he published on Wednesday, cautioning that AI safety protocols are lagging behind the swift evolution of AI, and asserting that the government “should be vested with the authority to halt or discourage the rollout of the model if, following independent evaluation, it is deemed to pose unacceptable risks.”
In a proposal accompanying his essay, Amodei announced a significant investment of $200 million dedicated to researching the societal ramifications of AI.
The Anthropic emblem featured in this visual, dated June 5, 2026.Dado Ruvic/Reuters
Regarding the impact on employment, Amodei informed ABC News that while the job market has historically adapted to technological advancements, this instance might necessitate intervention to support the public in both the short and long term.
“There are concepts like wage insurance, incentives for employee retention to dissuade companies from layoffs, and enhancements to unemployment benefits. These are measures to assist individuals through what may be a challenging but essentially normal transition,” he explained.
“In the interim, we need to discuss mechanisms for longer-term economic support, such as potentially increasing the capital gains tax or imposing taxes on AI corporations or corporate taxes generally,” he added.
Dario Amodei, CEO of Anthropic, gestures while speaking at a session on AI during the World Economic Forum (WEF) annual meeting in Davos on January 23, 2025.Fabrice Coffrini/AFP via Getty Images, FILE
When questioned about a recent statement from Anthropic suggesting a pause due to the technology’s near-autonomous development capabilities, Amodei indicated that a global agreement among nations and AI organizations would be essential for such a pause to be effective.
“I wish to clarify,” Amodei stated. “My viewpoint is that we should at least be prepared to attempt it.”
“It would necessitate the consensus of all pertinent AI entities, including geopolitical rivals, and it would need to be verifiable,” he elaborated.
Amodei expressed his skepticism regarding China’s potential for responsible conduct during an AI pause.
“I have absolutely no confidence in China,” Amodei declared.
“Consider if China had developed Mythos,” Amodei posited, referencing the advanced language model released by Anthropic earlier this year, designed to identify software vulnerabilities.
“They would have employed it for offensive purposes against us. They might have utilized it to assist Russia in Ukraine. Possibly, they would have aided Iran. It’s conceivable they would have leveraged it to support North Korea. One can only speculate about their potential actions with it,” he added.
During its evaluation phase, Claude Mythos demonstrated the capacity to exploit numerous flaws and coding deficiencies in prominent online operating systems and browsers, yet it was ultimately not made widely accessible.
This initiative, Amodei now contends, has reinforced his conviction that AI models require global regulation.
Despite these concerns, Amodei maintains an optimistic outlook regarding AI’s potential to benefit humanity, including its capacity to find cures for medical conditions.
“I believe AI holds the potential to eradicate numerous diseases that have afflicted humankind for centuries, if not millennia,” Amodei stated.
“I was previously involved in biology. Biological systems are extraordinarily intricate,” he continued. “AI possesses a remarkable aptitude for identifying connections within a vast array of diverse data, which I believe is precisely what is needed for advancements in biomedicine.”
Sourse: abcnews.go.com