Jack Clark, co-founder and Head of Public Benefit at Anthropic, has revealed that the AI firm briefed the Trump administration about its recently announced Mythos model. This advanced model, which includes significant cybersecurity capabilities, is deemed too risky for public release.
During an interview at Semafor’s World Economy Summit, Clark addressed Anthropic’s ongoing engagement with the U.S. government, despite the company’s lawsuit against the Department of Defense (DOD). Earlier this year, Anthropic contested the DOD’s classification of its technology as a supply chain risk, stemming from a conflict over military access to its AI systems for uses such as mass surveillance and fully autonomous weaponry. Clark attempted to downplay the DOD’s designation, characterising it as a minor contracting issue and asserting that national security remains a priority for Anthropic.
In his discussion, Clark emphasised the importance of collaboration between government and private sector entities, particularly regarding technological advancements that could impact national security. He confirmed that the company would continue to provide updates about Mythos and future models to the government.
Clark’s comments come in light of reports that the Trump administration is encouraging major banks, including JPMorgan Chase and Goldman Sachs, to test the Mythos model. He also touched on AI’s broader societal implications, particularly concerning employment and education. While Anthropic CEO Dario Amodei has expressed concerns that AI advancements could lead to substantial job losses, Clark offers a more tempered perspective, noting current weaknesses primarily in early graduate employment in certain sectors. Nonetheless, he acknowledges that the company is prepared for potential shifts in the job market.
When asked about academic pathways for students in response to AI’s influence, Clark advised pursuing interdisciplinary fields that foster analytical thinking and synthesis of diverse subjects. He argued that AI enables access to vast amounts of expertise across various domains, but success will hinge on the ability to ask relevant questions and derive insights from different disciplines.
Fanpage: TechArena.au
Watch more about AI – Artificial Intelligence
