Excessively Influential for the Public: A Look into Anthropic’s Efforts to Dominate the AI Public Relations Battle

In recent developments, the AI company Anthropic has announced the creation of a groundbreaking AI model so significant that they have chosen, out of a profound sense of responsibility, to withhold its public release.

Amidst this revelation, US Treasury Secretary Scott Bessent convened leaders from major banking institutions to discuss the implications of the model, named Mythos. Meanwhile, UK MP Danny Kruger reached out to the government, urging it to “collaborate with AI firm Anthropic, as the new model Claude Mythos raises potential cybersecurity threats in the UK.” Such commentary stirred considerable reactions across social media platforms.

Contrarily, some voices expressed skepticism, notably AI critic Gary Marcus. He stated that “Dario [Amodei] possesses more technical expertise than Sam [Altman] but appears to exhibit parallels in hype and exaggeration,” referring to the CEOs of Anthropic and its competitor, OpenAI.

While questions remain about whether Anthropic has indeed created a “machine god,” it is evident that this San Francisco startup excels in its marketing strategies.

Over the past few months, Anthropic has gained substantial media attention, featuring a lengthy profile in the New Yorker, multiple articles in the Wall Street Journal, and even landing the cover of Time magazine. Their co-founder Dario Amodei was showcased in a movie-poster style photograph prominently alongside the Pentagon and US Defense Secretary Pete Hegseth.

In February, Amodei and co-founder Jack Clark participated in two separate New York Times podcasts, discussing profound topics, including whether their AI is conscious and its potential impact on the economy. Additionally, the company’s “resident philosopher” has contributed to WSJ discussions about whether Claude—a tool utilized for trading cryptocurrency and identifying missile targets—possesses a “sense of self.”

Amidst these discussions, a notable tension has arisen between Anthropic and the US Department of Defense. In contrast to OpenAI, which offered assistance to the military, Anthropic has successfully positioned itself as a more trustworthy partner, even while being responsible for providing the AI tools used by the Pentagon for airstrikes.

Danielle Ghiglieri, the media lead for Anthropic, celebrated these achievements on LinkedIn. She expressed pride over Anthropic’s placement on the cover of Time, tagging the journalists involved in the endeavor while recalling the fast-paced efforts required to finalize the story.

Ghiglieri described watching a CBS 60 Minutes segment featuring Amodei as a “pinch-me moment,” emphasizing that the significance lay not only in the platform but also in narrating the story they aimed to share.

Reflecting on the New Yorker profile by Gideon Lewis-Kraus, Ghiglieri mentioned her initial nerves about collaborating with someone of his caliber, acknowledging the challenge of articulating ideas still in development.

Interestingly, this burgeoning media presence hasn’t gone unnoticed by peers in the tech PR field.

One expert pointed out that while Anthropic currently enjoys a favorable spotlight, it’s crucial for groundbreaking technology firms to undergo rigorous scrutiny. They noted that Anthropic had recently accidentally leaked part of Claude’s source code, which was a significant oversight. However, the company ensured that no sensitive customer data was compromised.

As we ponder the implications of Mythos, it’s imperative to scrutinize the validity of Anthropic’s assertions.

According to Dr. Heidy Khlaaf, chief AI scientist at AI Now Institute, the capabilities of the model have not been convincingly supported. She expressed that “promoting a marketing post with intentionally vague language raises suspicions regarding their intent to attract investment while avoiding oversight.”

“The development of Mythos is indeed significant and warranted a serious approach,” affirmed Jameison O’Reilly, an expert in offensive cybersecurity. However, he noted that some claims made by Anthropic, such as discovering numerous “zero-day vulnerabilities” in major operating systems, don’t hold substantial weight in practical cybersecurity contexts.

Zero-day vulnerabilities refer to defects in software or hardware that developers are unaware of when they’re exploited.

O’Reilly added, “Over the past decade, our team has gained authorized access to countless organizations—ranging from banks to governmental bodies. In that span of time, the number of occasions where we relied on a zero-day vulnerability to achieve our goals has been minimal.”

Various factors may have influenced Anthropic’s choice to withhold Mythos from release.

The company may be operating under limited resources and currently faces challenges in providing enough computational capacity for its subscriber base. Recently, Anthropic implemented usage caps on its popular model, Claude, suggesting users may now need to purchase additional capacity to access third-party tools like OpenClaw. Therefore, releasing a highly publicized new model may not be feasible given their current infrastructure limitations.

Similar to OpenAI, Anthropic finds itself in a race to secure billions in funding while striving to identify a potentially lucrative market that could encompass individuals seeking AI chatbots for companionship or businesses seeking to automate roles traditionally held by humans.

Nonetheless, distinctions between these AI products often stem from subjective assessments, largely hinged on nebulous concepts such as “sense of self” or “soul”—attributes that shape perceptions of AI agents rather than tangible differences in functionality. The competition now lies in captivating audience interests.

“The announcement regarding Mythos serves as a strategic maneuver to signify their readiness for business,” remarked Khlaaf. She cautioned that Anthropic’s choice to limit external evaluations on their claims may ultimately lead to potential public mistrust.

There is a possibility that we might be witnessing a familiar trend reminiscent of OpenAI, where the narrative of safety is utilized as a public relations strategy aimed at instilling public confidence before prioritizing profitability, suggesting that “Anthropic’s promotional efforts have succeeded in obscuring this transition more effectively than its competitors.”