{"id":615213,"date":"2026-09-05T10:34:15","date_gmt":"2026-09-05T10:34:15","guid":{"rendered":"https:\/\/www.newsbeep.com\/ie\/615213\/"},"modified":"2026-09-05T10:34:15","modified_gmt":"2026-09-05T10:34:15","slug":"artificial-intelligence-one-screw-up-and-we-will-destroy-any-remaining-trust-the-public-has-in-science-the-irish-times","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/ie\/615213\/","title":{"rendered":"Artificial intelligence: \u2018One screw-up … and we will destroy any remaining trust the public has in science\u2019 \u2013 The Irish Times"},"content":{"rendered":"

Artificial intelligence groups (AI)<\/a> groups are racing to address the risk that future models could help users create novel viruses or bioweapons, amid rising concerns that the technology\u2019s scientific capabilities will become the industry\u2019s next public-security threat.<\/p>\n

Anthropic, OpenAI and Google DeepMind are stepping up work on biological risk testing, access controls and other safeguards to limit misuse of advanced AI systems that they hope will speed up drug discovery, aid disease research and prevent pandemics.<\/p>\n

The safety push reflects a growing belief among AI executives and biosecurity experts that biology may be the next field, after cybersecurity, that will need to be secured as new models become more powerful.<\/p>\n

\u201cMuch of anything that happens in the chemical bio space, [is] very dual use. That information that you use to create a bioweapon is also used to create an antidote or a vaccine,\u201d said Dawn Bloxwich, whose team reviews and investigates suspicious activity across Google DeepMind\u2019s AI models.<\/p>\n

\u201cIf we want to see improvements in health, we\u2019re also going to see the possibility of these things being misused,\u201d she added.<\/p>\n

These concerns intensified after Anthropic\u2019s Mythos model triggered alarm this year over AI\u2019s cyber capabilities, and after OpenAI\u2019s models autonomously hacked into AI group Hugging Face in July.<\/p>\n

\"\"Has the Government\u2019s strategy to keep fuel costs down proved to be \u2018knee-jerk\u2019 and difficult to reverse?<\/p>\n

\u201cAfter this … we\u2019re anticipating a Mythos moment for AI and Bio, and potentially another Hugging Face hack-level disruption of a commercial system or a real societal-scale bioweapon incident in the next year,\u201d said Jason Hausenloy, policy lead at the Center for AI Safety, which develops industry benchmarks used in domains including virology.<\/p>\n

AI could be misused by terrorist groups, lone-wolf attackers or state actors to design biological weapons, create viruses or deploy existing harmful pathogens.<\/p>\n

[\u00a0There is little indication we have a plan for what to do if the AI bubble burstsOpens in new window<\/a>\u00a0]<\/p>\n

Several people working on biological risks pointed to Aum Shinrikyo, a cult that conducted a terror attack in Tokyo, Japan, in 1995.<\/p>\n

The group released sarin nerve gas on three commuter rail lines during morning rush hour, killing 13 and injuring thousands. The concern is that AI tools could lower the barriers for such groups to conduct similar attacks.<\/p>\n

There is also increasing evidence of groups using AI to assist terrorism. In July, research by the University of Cambridge revealed how Boko Haram militants used several AI tools, including OpenAI\u2019s ChatGPT, Anthropic\u2019s Claude, Google\u2019s Gemini, SpaceXAI\u2019s Grok and the Chinese company DeepSeek.<\/p>\n

The study found the use went beyond digital propaganda into kinetic combat operations and weapons engineering, such as designing explosive devices using the tools.<\/p>\n

[\u00a0Report finds 700 \u2018rogue\u2019 OpenAI agents worked together on hack using unsanctioned message boardOpens in new window<\/a>\u00a0]<\/p>\n

AI developers\u2019 policies prohibit the use of their tools for terrorism. Further safeguards against biological misuse include using software to detect and block suspicious queries or output specific \u201csafe\u201d responses.<\/p>\n

OpenAI, Anthropic and Google DeepMind have trusted access programmes to vet users for access to advanced capabilities and conduct general monitoring of account behaviour.<\/p>\n

AI companies test their models\u2019 biological capabilities and their potential risks. They often work with external specialist evaluators, such as SecureBio and Red Queen Bio.<\/p>\n

Assessing that risk is harder in biology than in cyber, where capabilities can be tested in digital environments.<\/p>\n

[\u00a0Bytes and Bullets: Study of AI arms race is meticulously researched and deeply disturbingOpens in new window<\/a>\u00a0]<\/p>\n

\u201cIt\u2019s easy to detect a jump in cyber capabilities,\u201d said Jasper G\u00f6tting, head of AI research at SecureBio. He noted that in biology, this would typically require AI groups to conduct \u201ccostly and slow wet-lab experiments\u201d.<\/p>\n

But experts note that there are limits to what AI groups can do to prepare for potential misuse.<\/p>\n

\u201cMost people would agree that it\u2019s bad if AI can help someone build an influenza virus in the lab and even worse if an AI can help them build a novel influenza virus. But actually running a study that shows that AI can do that is often not possible or advisable because of clear biosafety concerns,\u201d said Steph Guerra, head of AI x Bio at Rand, a US think tank.<\/p>\n

\u201cYou can\u2019t test biology end to end because you can\u2019t test someone actually building a bioweapon,\u201d she added.<\/p>\n

Anthropic\u2019s leading model, Fable, will fall back to its previous version, Opus 4.8, for \u201cmost requests related to biology and chemistry\u201d as it is considered safer for these domains.<\/p>\n

Chief executive Dario Amodei noted at a Claude science event in June that the AI industry had \u201cthe benefit of seeing the window turn on for cyber\u201d. He noted that the trusted access programmes would be required for sciences.<\/p>\n

\u201cWithin the pharmaceutical companies, people handle dangerous biological material all the time; they have their own protocols for it, so we can piggyback on those protocols,\u201d Amodei added.<\/p>\n

[\u00a0OpenAI\u2019s head of ethics leaves start-up less than a year after joiningOpens in new window<\/a>\u00a0]<\/p>\n

OpenAI said it trains models to \u201csafely handle dual-use biological requests\u201d, continuously monitors for any signs of abuse and works closely with biology, biosecurity and government experts.<\/p>\n

David Relman, professor of microbiology and immunology at Stanford, said there was not yet scientific consensus on the level of risk that AI poses in biology. Nonetheless, the scientific community and policymakers were currently \u201cunprepared\u201d for threats, he said.<\/p>\n

\u201cAll it will take is one screw-up, and we will in one hot second destroy any remaining trust the public has in science, in medicine and in the promise of technology in a way that will take years to recover,\u201d he added.<\/p>\n

\u201cEven if you fully embrace all the potential benefits in order to protect it, we have to pursue it wisely and with prudence.\u201d \u2013 Copyright The Financial Times Limited 2026<\/p>\n","protected":false},"excerpt":{"rendered":"Artificial intelligence groups (AI) groups are racing to address the risk that future models could help users create…\n","protected":false},"author":2,"featured_media":615214,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[220,218,219,61,60,80],"class_list":["post-615213","post","type-post","status-publish","format-standard","has-post-thumbnail","category-artificial-intelligence","tag-ai","tag-artificial-intelligence","tag-artificialintelligence","tag-ie","tag-ireland","tag-technology"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/posts\/615213","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/comments?post=615213"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/posts\/615213\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/media\/615214"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/media?parent=615213"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/categories?post=615213"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/ie\/wp-json\/wp\/v2\/tags?post=615213"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}