{"id":593867,"date":"2026-04-10T02:56:18","date_gmt":"2026-04-10T02:56:18","guid":{"rendered":"https:\/\/www.newsbeep.com\/ca\/593867\/"},"modified":"2026-04-10T02:56:18","modified_gmt":"2026-04-10T02:56:18","slug":"anthropic-says-new-claude-mythos-ai-is-too-risky-for-public-use-national","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/ca\/593867\/","title":{"rendered":"Anthropic says new Claude Mythos AI is too risky for public use – National"},"content":{"rendered":"

AI developer Anthropic <\/a>says its latest Claude AI model is so powerful \u2014 and potentially dangerous \u2014 that it will not be available to the general public to use.<\/p>\n

\"\"<\/p>\n

Dubbed Claude Mythos<\/a>, the software is part of the Claude AI family, an artificial intelligence<\/a> model that can act like a chatbot and AI assistant, like ChatGPT and Google\u2019s Gemini.<\/p>\n

\u201cIt is a frontier AI model, and has capabilities in many areas\u2014including software engineering, reasoning, computer use, knowledge work, and assistance with research\u2014that are substantially beyond those of any model we have previously trained,\u201d Anthropic wrote in the preview\u2019s system card<\/a>.<\/p>\n

The system card also states that Claude Mythos \u201chas demonstrated powerful cybersecurity skills, which can be used for both defensive purposes (finding and fixing vulnerabilities in software code) and offensive purposes (designing sophisticated ways to exploit those vulnerabilities).\u201d<\/p>\n

Story continues below advertisement<\/p>\n

It is those capabilities that made Anthropic decide to not release the software to the general public.<\/p>\n

\t\t\t\t\t\t\"Click<\/p>\n

\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t2:27
\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\tBusiness Matters: Pentagon disputes bolsters Anthropic\u2019s reputation, raises questions about AI use in warfare\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p>\n

\u201cClaude Mythos\u2019s large increase in capabilities has led us to decide not to make it generally available. Instead, we are using it as part of a defensive cybersecurity program with a limited set of partners.\u201d<\/p>\n

Anthropic cites these partners as \u201corganizations that maintain important software infrastructure, under terms that restrict its uses to cybersecurity.\u201d<\/p>\n

It is these kinds of technologies that Branka Marijan, a senior researcher at Project Ploughshares, says should be monitored with caution.<\/p>\n

\u201cThe implications for cybersecurity and broader national security that they are flagging, I don\u2019t think that they\u2019re hypotheticals,\u201d she said. \u201cI do think there are actual concerns that we should be paying more attention to now.\u201d<\/p>\n

Story continues below advertisement<\/p>\n

\t\t\t\t\t\t\"Click<\/p>\n

\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t0:42
\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\tPentagon labels Anthropic a \u2018supply chain risk\u2019\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p>\n

\t\t\t\tWhy is Claude Mythos stirring concern?\n\t\t\t\t\t\t<\/p>\n

Daniel Escott, the CEO of Formic AI, said that Anthropic is \u201cchoosing consciously\u201d to not release Claude Mythos.<\/p>\n

\u201cTheir argument against releasing it from the general public is that the same systems and functionality and capability to protect infrastructure using this AI system could equally be used to attack the same infrastructure,\u201d he said.<\/p>\n

Story continues below advertisement<\/p>\n

However, he also said that he would make \u201cno mistake\u201d that \u201csomeone will have access to [Claude] Mythos.\u201d<\/p>\n

\u201cAnthropic is making their own choices on who they\u2019re willing to give access to this system for. But at the same time, I would imagine those partners are probably saying \u2018you\u2019re only allowed to sell to us,\u2019 perhaps a limited set of other entities, but they don\u2019t want everyone to have access to the same kinds of technology,\u201d he said.<\/p>\n

\t\t\t\t\t\"Get<\/p>\n

\t\t\t\t\tGet breaking National news<\/p>\n

Get breaking Canada news delivered to your inbox as it happens so you won’t miss a trending story.<\/p>\n

\u201cAnd if Anthropic isn\u2019t going to sell it to them, someone else will develop it and sell it.\u201d<\/p>\n

\t\t\t\t\t\t\"Click<\/p>\n

\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t2:01
\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\tAnthropic vs Pentagon: Why the AI firm is pushing back against the Trump administration\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p>\n

Escott also warned that Anthropic\u2019s system card on Claude Mythos should be taken \u201cwith a grain of salt.\u201d<\/p>\n

\u201cBased on the documentation, it seems that they\u2019ve been training this on a combination of the open-source data sets that they\u2019d been using for all of Anthropic\u2019s other models,\u201d he said.<\/p>\n

Story continues below advertisement<\/p>\n

\u201cThis is no different than what ChatGPT or Microsoft Co-Pilot is doing, where they\u2019re just scraping, some would argue stealing, information from all over the internet and putting it all into one big data set that they can train on.\u201d\u00a0<\/p>\n

Marijan said she would like to see \u201cmore clarity from Anthropic\u00a0and these other companies about actually how concerning is this from what they\u2019re telling us.\u201d<\/p>\n

\u201cIt is absolutely concerning,\u201d she said.\u00a0\u201cIt\u2019s undermining all of these safeguards that companies might have in place.\u201d<\/p>\n

\t\t\t\tWould Anthropic ever release Claude Mythos?\n\t\t\t\t\t\t<\/p>\n

Moshe Lander, an economics professor at Concordia University, said that not releasing Claude Mythos to the public just yet allows for potential flaws to be fixed without impacting users.<\/p>\n

\u201cIf some pharmaceutical company is developing a drug, and they say, for the time being, \u2018we\u2019re not releasing it for public use,\u2019 is there something wrong with that? I would say, actually, I think that\u2019s probably being responsible,\u201d he said.<\/p>\n

Story continues below advertisement<\/p>\n

\u201cIf the company is saying, \u2018look, we\u2019re not putting it into public use ever,\u2019 that\u2019s something different. What they\u2019re saying is \u2018we\u2019re now putting it in public use now,\u2019 I think that\u2019s being extremely responsible, in let\u2019s see how this thing is going to be used. Let\u2019s see where its defects are,\u201d he said.<\/p>\n

\u201cIf they do find that there\u2019s weaknesses, it has that ability to correct itself or fix any flaws, that might not be a bad thing.\u201d<\/p>\n

\t\t\t\t\t\t\"Click<\/p>\n

\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t1:50
\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\tFederal government looks to support Canadian AI industry\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p>\n

There remain significant questions around the world, including in Canada, around what it will take for governments to regulate AI and provide legal frameworks for its use.<\/p>\n

Lander also said that initial concern about AI systems not being immediately released is bound to raise questions for many, with no easy answers.<\/p>\n

Story continues below advertisement<\/p>\n

\u201cI think that because people are generally worried about AI in general, that when we hear there\u2019s an AI product that\u2019s coming along that\u2019s not available for public use, we hit the panic button and say, \u2018wait a second, something doesn\u2019t sound right here,’\u201d he said.<\/p>\n

\u201cBefore they [Anthropic] put it into public use, they want to make sure that it\u2019s not going to go into the wrong hands, where people have maybe dishonourable intentions and that it can be used to harm society once they\u2019ve established the protocols or safeguards that we need to put in place.\u201d<\/p>\n

\t\t\t\tRansomware attacks increasing for Canadian businesses\n\t\t\t\t\t\t<\/p>\n

In January<\/a>, the Canadian Centre for Cyber Security (Cyber Centre) released its ransomware threat outlook for 2025-27, <\/a>stating that with the growth of AI, \u201cthese threats have become cheaper and faster to conduct and harder to detect.\u201d<\/p>\n

Story continues below advertisement<\/p>\n

As a result, numerous Canadian organizations, businesses \u201cregardless of size or sector,\u201d and individuals are susceptible to ransomware attacks. However, \u201ccritical infrastructure and large corporations\u201d were found to be the top targets for ransomware activities.<\/p>\n

\t\t\t\t\t\t\"Click<\/p>\n

\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t1:29
\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\tCanadian firms paying \u2018significantly\u2019 more in ransomware attacks: data\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/p>\n

The report found that the reported number of ransomware incidents increased by an average of 26 per cent year over year from 2021 to 2024.<\/p>\n

In addition, it was also found that the total recovery costs associated with cybersecurity incidents cost $1.2 billion in 2023, doubling the previous cost of $200 million from 2019 to 2021.<\/p>\n

However, Marijan believes that there should be more protocol in place for businesses to utilize these tools.<\/p>\n

\u201cI think what it points to really is this clear gap in governance where we have companies that are deciding what they think is concerning. We should really have processes,\u201d she said.\u00a0<\/p>\n

Story continues below advertisement<\/p>\n

\u201cWhat we\u2019ve seen over the last decade is an increase in ransomware attacks [\u2026] and that impacts all of us. So, when you\u2019re thinking about \u2018what are the implications of these,\u2019 they\u2019re very significant for ordinary people as well.<\/p>\n

\u201cSo, we absolutely are in the space where these companies are deciding essentially what they think are concerns or flagging them. And there\u2019s no process in place for this, for any guardrails really to appear.\u201d<\/p>\n","protected":false},"excerpt":{"rendered":"AI developer Anthropic says its latest Claude AI model is so powerful \u2014 and potentially dangerous \u2014 that…\n","protected":false},"author":2,"featured_media":593868,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[5],"tags":[1930,276,45,49,48,225716,64],"class_list":{"0":"post-593867","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-business","8":"tag-anthropic","9":"tag-artificial-intelligence","10":"tag-business","11":"tag-ca","12":"tag-canada","13":"tag-claude-mythos","14":"tag-tech"},"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/posts\/593867","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/comments?post=593867"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/posts\/593867\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/media\/593868"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/media?parent=593867"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/categories?post=593867"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/tags?post=593867"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}