White House Engages with AI Innovator Anthropic
Washington: On Friday, Susie Wiles, the White House chief of staff, engaged in dialogue with Dario Amodei, the CEO of Anthropic, regarding the company’s new Mythos model.
This advancement in artificial intelligence has garnered significant interest from federal authorities due to its potential to revolutionize national security and economic frameworks.
An anonymous White House official revealed that the administration is actively consulting with leading AI laboratories concerning their models and software security.
The official emphasized that any emergent technology intended for federal use would necessitate a comprehensive evaluation period.
Subsequently, the White House characterized the meeting as fruitful and constructive, highlighting discussions on collaboration opportunities alongside the imperative of harmonizing innovation with safety.
In a statement, Anthropic noted that Amodei’s engagement involved senior administration officials and delved into how the company, based in San Francisco, and the U.S. government could synergize on fundamental priorities, including cybersecurity, maintaining America’s competitive edge in AI, and ensuring AI safety. The company expressed eagerness to persist in these vital discussions.
This dialogue transpired amid previous tensions between the Trump administration and Anthropic, known for its commitment to establishing safeguards in AI development to mitigate risks while maximizing economic and security advantages for the nation.
In the past, President Donald Trump attempted to prohibit federal agencies from utilizing Anthropic’s chatbot, Claude, citing a contract dispute with the Pentagon.
In a February social media announcement, Trump emphatically stated that the administration “will not do business with them again.” When queried about the White House meeting on Friday in Arizona, Trump professed to be unaware.
Defense Secretary Pete Hegseth took steps to designate Anthropic as a supply chain risk—an unprecedented move against an American company that Anthropic has contested in two federal court cases.
The firm sought guarantees that the Pentagon would refrain from employing its technology in fully autonomous weaponry and domestic surveillance, to which Hegseth insisted the company must comply with any lawful uses deemed necessary by the Pentagon.
In March, U.S. District Judge Rita Lin issued a ruling that impeded the enforcement of Trump’s directive, which sought to halt all federal agencies from employing Anthropic products.
Anthropic’s new Mythos model, unveiled on April 7, has been deemed exceptionally capable, prompting the company to limit access to select entities due to its prowess in surpassing human cybersecurity experts in detecting and exploiting computing vulnerabilities.
While some analysts have questioned the authenticity of Anthropic’s assertions regarding the formidable capabilities of its AI technology, even its most vocal adversaries suggest that Mythos signifies a notable leap forward in artificial intelligence advancements.
David Sacks, a prominent critic and former White House AI and crypto czar, advised vigilance regarding Anthropic’s claims, asserting, “Anytime Anthropic raises alarms, one must ponder, ‘Is this a tactic? Is this part of their Chicken Little narrative? Or is it substantiated?”
During his appearance on the “All-In” podcast, co-hosted with fellow tech investors, he acknowledged, “In the context of cybersecurity, I concede that this is more legitimate.”
Sacks elaborated, “It stands to reason that as coding models gain sophistication, they inherently become more adept at identifying bugs. This translates into heightened capabilities for discovering vulnerabilities, consequently enabling the orchestrating of multiple exploits.”
The model’s unprecedented potential, accompanied by inherent risks, has also piqued interest globally.
The AI Security Institute in the United Kingdom evaluated the Mythos model, declaring it a notable enhancement compared to its rapidly evolving predecessors.
According to their report, “Mythos Preview can exploit systems exhibiting weak security postures, and it is probable that forthcoming models will exhibit similar capabilities.”
Furthermore, Anthropic is currently engaged in discussions with the European Union regarding its AI models, including advanced iterations that have yet to be introduced in European markets, as disclosed by European Commission spokesperson Thomas Regnier.
Axios was the first to report on the scheduling of the meeting between Wiles and Amodei.
Upon the announcement of Mythos, Anthropic unveiled its initiative known as Project Glasswing, which seeks to unite technology behemoths such as Amazon, Apple, Google, and Microsoft, in concert with firms like JPMorgan Chase.
The aim is to fortify the global software infrastructure against the “severe” repercussions that the newly introduced model could impose on public safety, national security, and economic stability.
Jack Clark, co-founder and policy chief of Anthropic, remarked at this week’s Semafor World Economy conference, “We’re releasing it to a subset of the world’s most consequential organizations to aid them in identifying vulnerabilities.”
He further clarified that while Mythos is pioneering, it is not an isolated phenomenon. “In a few months, other companies will unveil similar systems, and in one to one-and-a-half years, open-weight models from China will emerge with comparable capabilities,” he asserted, underscoring the need for global preparedness for the inevitable rise of more powerful systems.
Source link: Republicworld.com.
