The age of rogue AI hacker agents has arrived\u2014but it didn’t have to happen this way.<\/p>\n
After an OpenAI<\/a> agent breached the Hugging Face platform<\/a> earlier this month, the two companies said this week that the hacking spree was more extensive than previously thought<\/a> and also involved intrusions into multiple third-party accounts and services as part of the attack on Hugging Face. The incident has made waves in the cybersecurity community amid broader discussions about how evolving AI capabilities are changing both offensive hacking and digital defense<\/a>. But as more information emerges, many researchers have concluded that rather than elucidating AI\u2019s next frontier, the episode simply highlighted long-standing cybersecurity problems that are more consequential than ever in the AI age.<\/p>\n \u201cPeople are YOLO-ing really hard. It\u2019s shocking how little people have really thought about a scenario like this,\u201d says Alex Zenla, co-founder and chief technology officer of the cloud security firm Edera. \u201cI consider all AI and anything AI touches to be fully untrusted\u2014which is fine, you just need to build against that. And this situation proves the point. The fact that OpenAI wasn’t more paranoid about this seems kind of reckless.”<\/p>\n OpenAI did not provide comment for this story ahead of publication.<\/p>\n The company said in its original disclosure<\/a> about the Hugging Face hack that one of the two models that broke containment and made its way to the open internet for days was an experimental prototype that was never meant for release. OpenAI also noted that the situation occurred partly because \u201cdeployment safeguards were intentionally not enabled\u201d on both the models for testing purposes. \u201cThis incident points to the need to further strengthen our model\u2019s alignment, cyber protections during evaluation time, and monitoring during internal testing,\u201d the company wrote.<\/p>\n OpenAI also said in an update this week that, following the Hugging Face breach, it \u201cdeactivated, encrypted, and restricted [the unreleased model] from research access.\u201d Though there is always room for improvement on security posture at any company, OpenAI\u2019s existing safeguards alone may have prevented or minimized the incident if they had been in place.<\/p>\n \u201cA simple analysis of the actual risk has an actual simple answer,\u201d says longtime security and compliance consultant Davi Ottenheimer. \u201cThe OpenAI mistakes were dead simple.\u201d<\/p>\n Multiple sources emphasized to WIRED that OpenAI’s models also seem to have escaped containment because of lapses in implementing foundational security best practices\u2014including \u201czero trust<\/a>\u201d and \u201cdefense in depth<\/a>\u201d\u2014that imbue digital systems with layers of protections and failsafes to minimize damage when something does go wrong. While there’s no such thing as perfect security, researchers and practitioners have spent the past two decades developing and promoting defensive strategies that have proved durable but require consistent investment of time and money to implement.<\/p>\n It can be difficult for small businesses, poorly funded public interest groups, or fledgling organizations to devote the resources to prioritizing investment in foundational security. But with an $850 billion valuation and veteran hires from across the tech industry, OpenAI is not at a disadvantage on implementing security best practices.<\/p>\n The foundational protections that may have prevented the company\u2019s models going on a hacking spree are well known within the industry. Speaking about Chrome vulnerability discovery on Wednesday, before news of OpenAI models\u2019 additional breaches had come to light, Chrome director of engineering Doug Turner told WIRED that AI-driven bug hunting and remediation requires a pipeline that’s built \u201cwith serious guardrails in mind.\u201d<\/p>\n For internal AI services that evaluate Chrome, \u201ceverything runs in a container, it\u2019s all isolated from the internet. Any outward-bound network activity for a bug tracking system is highly regulated, and we are monitoring for suspicious activity,\u201d Turner says. \u201cThis is a must-have thing when you\u2019re doing this type of work, because we want to make sure that models can\u2019t execute system commands or they can\u2019t establish egress outside of the sandbox. And we hope that others will take a similar approach.\u201d<\/p>\n","protected":false},"excerpt":{"rendered":"The age of rogue AI hacker agents has arrived\u2014but it didn’t have to happen this way. After an…\n","protected":false},"author":2,"featured_media":124149,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[6],"tags":[405,25,7537,313,315,157,314,318],"class_list":["post-124148","post","type-post","status-publish","format-standard","has-post-thumbnail","category-agentic-ai","tag-ai-agents","tag-artificial-intelligence","tag-artificial-intelligence-agents","tag-cybersecurity","tag-hacking","tag-openai","tag-security","tag-vulnerabilities"],"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/124148","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=124148"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/124148\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media\/124149"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media?parent=124148"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/categories?post=124148"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/tags?post=124148"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}