Jamie Dimon warns Anthropic Mythos AI poses real risks as government tightens oversight

JPMorgan CEO Dimon says Anthropic's Mythos AI risks are a 'real issue'

July 15 (Reuters) - JPMorgan Chase (JPM.N), opens new tab CEO Jamie Dimon said on Wednesday that risks posed by Anthropic's Mythos AI model are a "real issue" that the U.S. government is on top of, underscoring concerns about the powerful model that have caused Washington to intervene. Dimon, speaking at Senator Dave McCormick's Pennsylvania Defense and Innovation Summit, added that access to advanced AI capabilities must be controlled, saying "you're giving ballistic ⁠missiles to individuals with Mythos." U.S.-based Anthropic released Mythos in April to a select group that included JPMorgan. The model has attracted interest from banks because it is considered among the most advanced AI systems for identifying cybersecurity vulnerabilities, helping firms uncover and fix weaknesses more quickly. However, in June, the U.S. government ordered Anthropic to restrict access to its top AI models, Fable 5 and ⁠Mythos 5, to foreign nationals over national security concerns, forcing the company to cut off worldwide access. The government later removed restrictions on the models after Anthropic implemented new safeguards. Governments and tech companies are ⁠grappling with AI safety concerns, including the misuse of systems capable of identifying software vulnerabilities. Washington has stepped up oversight of new model ⁠releases amid fears that advanced AI could be exploited by military intelligence in China, Russia or other countries of ⁠concern. Anthropic could not be immediately reached for comment on Dimon's remarks outside regular business hours. Reporting by Mihika Sharma in Bengaluru and Mrinmay Dey in Mexico; Editing by Rashmi Aich Our Standards: The Thomson Reuters Trust Principles., opens new tab

Jamie Dimon warns about Anthropic Mythos AI access

JPMorgan $JPM Chase & Co. Chief Executive Officer Jamie Dimon described the risks of broad public access to Anthropic PBC's Mythos artificial intelligence model as a "real issue," comparing it to handing ballistic missiles to individuals. He made the remarks Wednesday at Senator Dave McCormick's Pennsylvania Defense and Innovation Summit. "You're giving ballistic missiles to individuals with Mythos, basically," Dimon said. He added that the U.S. government is on top of the issue now. The warning centers on Mythos's capacity to identify software vulnerabilities at a level Anthropic has said is too dangerous for general release. JPMorgan is among a select group of companies that gained access to the model in April and has been deploying it to stress-test internal systems and coordinate on findings with vendors and peers. The month after gaining that access, Dimon said the bank had hundreds of employees working full time to strengthen its systems. JPMorgan's involvement with Mythos is part of its broader push into national security and economic resilience. The bank announced last year a "Security and Resiliency Initiative" through which it would direct $1.5 trillion over ten years toward sectors that strengthen the U.S. economy -- a figure 50% above what it would have invested otherwise. JPMorgan also tied the commitment to its separate "American Dream Initiative," pledging $24 million in a mix of loans, investments, and philanthropic grants directed at Philadelphia's shipbuilding sector. Dimon visited the Philadelphia Navy Yard earlier Wednesday as the bank announced that commitment. The Mythos model has been at the center of a broader government and industry reckoning over AI safety. The U.S. government lifted export controls on Anthropic's Claude Fable 5 and Mythos 5 models on July 1 after an 18-day shutdown triggered by Amazon $AMZN researchers documenting a technique for eliciting dangerous outputs from Fable 5 related to software vulnerability discovery. A limited group of U.S. organizations regained access to Mythos 5 after the government gave its approval on June 26; extending that access to additional domestic and international partners through Anthropic's Glasswing program remains under discussion with federal authorities, Anthropic said. Dimon has grown increasingly vocal on national security matters in recent years. His April shareholder letter argued that America must "get stronger" if it hopes to preserve its position as a military and economic power.

Anthropic's Mythos AI: JPMorgan CEO Jamie Dimon says Anthropic's Mythos AI risks are a 'real issue'

Dimon, speaking at ⁠Senator ⁠Dave McCormick's Pennsylvania Defense and Innovation Summit, added that access to advanced AI capabilities must be controlled, saying "you're giving ballistic missiles to individuals with Mythos." JPMorgan Chase CEO Jamie Dimon said on Wednesday that risks posed by Anthropic's Mythos AI model are a "real issue" that the U.S. government is on top of, underscoring concerns about the powerful model that have caused Washington to intervene. Dimon, speaking at ⁠Senator ⁠Dave McCormick's Pennsylvania Defense and Innovation Summit, added that access to advanced AI capabilities must be controlled, saying "you're giving ballistic missiles to individuals with Mythos." U.S.-based Anthropic released Mythos in April to a select group that included JPMorgan. The model has attracted interest from banks ⁠because it is considered among the most advanced AI systems for identifying cybersecurity vulnerabilities, helping firms uncover ⁠and fix weaknesses more quickly. However, in June, the U.S. government ordered Anthropic to restrict access to its top AI models, Fable 5 and Mythos 5, to foreign nationals over national security concerns, forcing the company to cut off worldwide access. The government later removed restrictions on the models after Anthropic implemented new safeguards. Governments and tech companies are grappling with AI safety concerns, including the misuse of systems capable ⁠of identifying software vulnerabilities. Washington has stepped up oversight of new model releases amid fears that advanced AI could be exploited by military intelligence in China, Russia or other countries of concern. Anthropic could not be immediately reached for comment on Dimon's remarks outside regular business hours.

Jamie Dimon Says Anthropic's Mythos Access Debate Is a Warning Sign for AI's Future - JPMorgan Chase (NYS

Anthropic has found itself at the center of debate over how governments should manage cybersecurity risks posed by powerful AI models. Anthropic has since added more companies to the project, expanding its access to approximately 150 new organizations, the company noted last month. "Project Glasswing and the capabilities of Claude Mythos Preview have sparked broad conversations -- both within the software industry and with governments -- about how AI is changing cybersecurity," Anthropic said. "We see our role as twofold. First, to help the software industry adapt by safely providing wide access to better models, tools and common infrastructure. Second, to steadily shift the support we provide, from finding vulnerabilities to disclosing, fixing, and deploying patched software," the company stated at the time. Founded in 2021 by former OpenAI employees, Anthropic has quickly established itself as a key player in artificial intelligence, building tools that are reshaping how businesses approach areas like coding and cybersecurity. In May, Anthropic overtook OpenAI as the world's most valuable startup after raising $65 billion in Series H, valuing the company at $965 billion. Altimeter Capital, Dragoneer, Greenoaks and Sequoia Capital led the funding round. Anthropic is meeting with investors ahead of a potential blockbuster IPO as it prepares to capitalize on strong demand for AI listings. Banks leading the offering are arranging investor meetings in the coming weeks, as the company is reportedly targeting an IPO as early as October. The Claude maker has hired Morgan Stanley, Goldman Sachs Group and JPMorgan Chase to work on the IPO process. Details regarding the IPO are ongoing and subject to change. Market News and Data brought to you by Benzinga APIs To add Benzinga News as your preferred source on Google, click here.