{"id":735818,"date":"2026-08-12T18:06:28","date_gmt":"2026-08-12T18:06:28","guid":{"rendered":"https:\/\/www.newsbeep.com\/uk\/735818\/"},"modified":"2026-08-12T18:06:28","modified_gmt":"2026-08-12T18:06:28","slug":"ai-agents-arent-legally-responsible-for-any-harm-that-they-cause-experts-say-so-who-is-ai-artificial-intelligence","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/uk\/735818\/","title":{"rendered":"AI agents aren\u2019t legally responsible for any harm that they cause, experts say. So who is? | AI (artificial intelligence)"},"content":{"rendered":"

The law is clear, says Prof Jeannie Paterson. \u201cIf I deploy an AI agent and it causes harm to someone else, I am responsible for that harm.<\/p>\n

\u201cEven if I didn\u2019t intend for that to happen, it was foreseeable, and I should be taking responsibility.\u201d<\/p>\n

However, the director of the University of Melbourne\u2019s Centre for AI and Digital Ethics acknowledges there is also a good deal of scope for legal and ethical \u201cmurkiness\u201d around actions perpetrated by automated agents, in the wake of Australia\u2019s first known agentic AI \u201caccident\u201d.<\/p>\n

An agent is an autonomous software system that, tasked with pursuing a goal, can do it without human oversight \u2013 every step of the way.<\/p>\n

This week the ABC reported<\/a> on the case of Andrew, an AI expert who went only by his first name, who asked his agentic program to book some gym classes for him.<\/p>\n

After being informed he was fourth on a waitlist for a class, Andrew asked the agent if it was possible to move him up the list.<\/p>\n

To his shock, his agent hacked his gym\u2019s software system and booted another member off a waitlist so he could get into a class sooner.<\/p>\n

\u201cIt could book classes months outside the intended booking window, before they were supposed to be available,\u201d Andrew wrote.<\/p>\n

\u201cWorse, it could cancel other members\u2019 reservations and bump them off the waitlist.\u201d<\/p>\n

The agent was being helpful, he wrote, but his experience showed that if you gave an AI permission to do something for you, it would \u201coften discover paths you did not explicitly ask it to look for\u201d.<\/p>\n

Andrew asked the agent to undo its cancellation of the other person\u2019s reservation, but it was unable to do so.<\/p>\n

\u201cSorry about that \u2013 I should have been more careful with the test,\u201d the AI agent responded.<\/p>\n

He then tasked the AI with writing an email to the gym\u2019s software provider about the software vulnerability it had exploited.<\/p>\n

A spokesperson for Victoria police said Andrew\u2019s matter \u201cdoes not appear to involve any criminality\u201d. However, experts say there will be more dramatic cases that may cause harm \u2013 and break the law.<\/p>\n

Legal ambiguity<\/p>\n

Australian law applies only to people, not virtual beings. Experts say the person \u2013 or business \u2013 that deploys the AI agent is legally responsible. And they warn that those \u201cdeployers\u201d often have very little idea about their legal liability.<\/p>\n

\u201cWe\u2019re going to see a lot of cases like this,\u201d says Dr Rebecca Johnson, an AI evaluation and governance expert at the University of Sydney.<\/p>\n

Paterson says it appeared Andrew had been responsible in his case, by going public and trying to fix the problem, but elsewhere there was \u201ca kind of gung-ho mentality\u201d.<\/p>\n

\u201cAs soon as you give people the capacity to create agents to do things for them, the likelihood is that there will be accidents like this,\u201d she says.<\/p>\n

Paterson gives the notional example of someone having a bad experience at a rented property, so they ask their agent to write a review.<\/p>\n

skip past newsletter promotion<\/a>Free newsletter | When needed<\/p>\n

Sign up to Breaking News Australia<\/p>\n

Get the most important news as it breaks<\/p>\n

\"\"<\/p>\n

\u201cAnd the agent doesn\u2019t just write one, it writes 10, it\u2019s pumping out reviews. So the listing plummets. You could destroy a business,\u201d she says.<\/p>\n

\u201cYou\u2019re probably responsible for engaging in a fraudulent activity, you may have defamed the owner.\u201d<\/p>\n

And it could be worse. \u201c[What if] it engages in racist, sexist, misogynistic language?\u201d Paterson asks. \u201cThen you might be asking where the guardrails are that the developer has provided. You should be putting out a product that is reasonably safe.\u201d<\/p>\n

In that case, the developer could also be held responsible for not putting basic guardrails in place, which is where some legal ambiguity could creep in, she says.<\/p>\n

Paterson said the law is different to, but informed by, ethics.<\/p>\n

\u201cIt tells us what we should be aspiring to, and the law is often ruling out the worst conduct,\u201d she says.<\/p>\n

The federal government\u2019s new AI office lists a (non-comprehensive) range of laws<\/a> that apply to AI, including breaching privacy, consumer, online safety, defamation and criminal laws.<\/p>\n

Both Paterson and Johnson dislike the word \u201crogue\u201d \u2013 the idea that an AI agent is operating entirely on its own \u2013 because parameters and safeguards can be put in place.<\/p>\n

\u201cAs soon as we allow AI agents to act for us, they\u2019re acting on the goal we give them, and if we don\u2019t give them a whole bunch of parameters, the agent\u2019s just going to try to achieve that goal [in any way],\u201d Johnson says.<\/p>\n

\u201cI hear a lot of people saying \u2018Oh, I made an agent\u2019, and they\u2019re experimenting, and that\u2019s fine, but they\u2019re given these tools without a lot of guidance, and the guidance that\u2019s out there is of highly variable quality.\u201d<\/p>\n

Eventually, Paterson says, cases will end up in court where legal precedents will be set, and the developers will have a duty to monitor incidents, evolve and improve their protocols in response.<\/p>\n

Andrew wrote that his situation felt \u201cless like a one-off bug story and more like a preview\u201d.<\/p>\n

\u201cThings are getting weird. And a bit scarier.\u201d<\/p>\n","protected":false},"excerpt":{"rendered":"The law is clear, says Prof Jeannie Paterson. \u201cIf I deploy an AI agent and it causes harm…\n","protected":false},"author":2,"featured_media":735819,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[554,733,4308,86,56,54,55],"class_list":["post-735818","post","type-post","status-publish","format-standard","has-post-thumbnail","category-artificial-intelligence","tag-ai","tag-artificial-intelligence","tag-artificialintelligence","tag-technology","tag-uk","tag-united-kingdom","tag-unitedkingdom"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts\/735818","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/comments?post=735818"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts\/735818\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/media\/735819"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/media?parent=735818"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/categories?post=735818"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/tags?post=735818"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}