![\"anthjropic-mythos-gs0429\"](\"https:\/\/www.europesays.com\/ai\/wp-content\/uploads\/2026\/04\/45b3a172cf722af45f69eb0c1ba269c3.jpeg\")
Earlier this month, Anthropic introduced Mythos, an AI model that it said excels at uncovering and exploiting software vulnerabilities. (Credit: SEBASTIEN BOZON) <\/p>\n Earlier this month, Anthropic introduced Mythos, an AI model that it said excels at uncovering and exploiting software vulnerabilities. (Credit: SEBASTIEN BOZON) <\/p>\n
The emergence of Anthropic PBC<\/a>\u2019s Mythos is a signal point for the financial system to \u201cput on our thinking caps and our glasses\u201d to understand what risks artificial intelligence<\/a> poses, says the country\u2019s top banking regulator.<\/p>\n The industry needs to \u201creally understand what\u2019s coming at us, because it\u2019s coming at us pretty fast,\u201d Peter Routledge, head of the Office of the Superintendent of Financial Institutions<\/a> (OSFI), said at an event on Tuesday.<\/p>\n He also said neither OSFI<\/a> nor banks can tackle those risks alone.<\/p>\n \u201cIt\u2019s got to be a shared effort,\u201d he said.<\/p>\n Earlier this month, Anthropic introduced Mythos<\/a>, an AI model that it said excels at uncovering and exploiting software vulnerabilities. The tool is supposedly so powerful that the company has declined to release it to the public. The fear is that if the tool falls in the wrong hands, there could be a massive increase in cyber risk.<\/p>\n Routledge\u2019s comments mark a slight shift in tone for OSFI. Over the past two years, the regulator focused on understanding AI rather than regulating it.<\/p>\n \u201cI sort of said, \u2018It\u2019s too early to jump into regulation and our first principle should be to do no harm and enable institutions to innovate with AI,’\u201d he said. \u201cWe approached (AI) with a healthy restraint.\u201d<\/p>\n Canada\u2019s financial sector, especially the Big Six banks<\/a>, has experimented with AI and looked for ways to use it to boost productivity.<\/p>\n For example, Canadian Imperial Bank of Commerce<\/a> said AI helped it save 1.2 million hours of work during its first quarter, while Toronto-Dominion Bank<\/a> said AI allowed it to cut reviews of its mortgage approval process to minutes from hours.<\/p>\n But the emergence of Mythos doesn\u2019t mean OSFI is going to \u201csurrender\u201d its \u201cdo-no-harm principle,\u201d Routledge said, but he wants financial institutions to be better aware of the potential risks of AI and Mythos.<\/p>\n \u201cHow serious Mythos is, how serious a threat it is, we still have a cycle to go through to truly understand that,\u201d he said.<\/p>\n Routledge said he wants board members and senior managers to meet quarterly to price in \u201cthe severity of the risk\u201d and develop defences. He also said there will have to be a public-private sector approach to tackle this risk.<\/p>\n \u201cIt\u2019s very conceptual,\u201d he said. \u201cWe just haven\u2019t gotten into the details of how to build those defences and what our role, or other institutions in the public sector, might have to do.\u201d<\/p>\n Still, Routledge said he is optimistic the sector can strengthen its safeguards in a \u201chealthy amount of time.\u201d<\/p>\n OSFI\u2019s annual risk outlook, which was released earlier this month, didn\u2019t focus on Mythos, but on the risks related to the residential housing market, non-banking financial institutions and liquidity, but that\u2019s because it was drafted in January.<\/p>\n","protected":false},"excerpt":{"rendered":"Earlier this month, Anthropic introduced Mythos, an AI model that it said excels at uncovering and exploiting software…\n","protected":false},"author":2,"featured_media":21606,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[8],"tags":[53,1616,353,15093],"class_list":["post-21605","post","type-post","status-publish","format-standard","has-post-thumbnail","category-anthropic","tag-anthropic","tag-financial-institutions","tag-mythos","tag-osfi"],"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/21605","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/comments?post=21605"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/21605\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media\/21606"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media?parent=21605"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/categories?post=21605"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/tags?post=21605"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}