AnthropicのMythosがAI研究所と企業に意味すること

3 Min ReadWith cybersecurity becoming an increasingly important area for AI, it's no surprise that Anthropic's upcoming model focuses on enterprise cybersecurity. However, businesses likely don't view the AI lab as the sole provider of all their AI needs.A data leak reported by Fortune on March 26 exposed details about a new Anthropic model called Claude Mythos. While the model maker has not confirmed the blog post to AI Business, it confirmed to other publications that it has completed training on a model called Mythos.According to the leaked blog, Mythos is Anthropic's most powerful and compute-intensive model, making it expensive. The model is also advanced at identifying and exploiting software vulnerabilities that could be a cybersecurity risk, according to the leaked document.Mythos' focus on cybersecurity led to a sharp decline in cybersecurity stocks on March 27. The response by cybersecurity investors mirrors that of SaaS and legal investors in February when Anthropic introduced Claude Cowork. This reaction is another example of fear among investors in traditional sectors about the extent to which AI can replace established technologies.Related:The Real AI Shift Isn’t New Models. It’s Control.Diversifying From CodingHowever, the upcoming model also shows how Anthropic continues to follow the various opportunities it sees in the application of AI, said Lian Jye Su, an analyst at Omdia, a division of Informa TechTarget."Making coding the only thing has become a bit of a stale thing for Anthropic," Su said. "Obviously, they need to expand their market as well. Being so targeted at cybersecurity is also quite interesting."He added that while it's unclear what performance Mythos will deliver, it's reasonable to expect it to perform better than the vendor's existing stable of models. Moreover, Anthropic has dabbled in security before with Claude Security Code, which it released in a limited preview in February. "Anthropic is trying to showcase to the world that they are serious about anything security-related, so very on-brand in a way and also signaling to the industry that they are a worthy partner when it comes to anything security," Su said.The Response and EnterprisesHowever, the reaction in the cybersecurity market is likely an overreaction, he said. He added that while cybersecurity startups may feel pressure that AI labs can develop a competitive cybersecurity model, established cybersecurity software vendors such as Palo Alto Networks and CrowdStrike might not share that concern."It's already a known fact that most cybersecurity companies will be using models from this foundation model provider," he said. "It's just [a question] of how much fine-tuning they perform on the model itself and how much they develop internally."Related:OpenAI GPT-5.4-Cyber is More Open Than Claude Mythos"It points more toward a future where AI now becomes a lot more embedded in various applications," Su continued.For enterprises watching all of this play out, the goal will be to look for a good AI partner, Su said. He added that given how complex cybersecurity is and the fact that companies are dealing with circumstances like "shadow AI environments, you are running everything distributed from cloud to the edge. There are various sorts of silos and unstructured systems."This means businesses need different types of tools, and while Anthropic can be one of them, it does not negate the importance of other tools and providers.About the AuthorNews Writer, AI BusinessEsther Shittu brings four years of expertise covering artificial intelligence technologies and industry trends. As co-host of the "Targeting AI" podcast, she talks to thought leaders and practitioners exploring critical AI developments. Previous to AI Business, she wrote for several publications including the New York Daily News, Bklyner and the Brooklyn Daily Eagle. When she's not diving deep into the world of AI, she spends her time on passion projects and raising her three daughters.