{"id":421032,"date":"2026-01-21T15:32:11","date_gmt":"2026-01-21T15:32:11","guid":{"rendered":"https:\/\/www.newsbeep.com\/us\/421032\/"},"modified":"2026-01-21T15:32:11","modified_gmt":"2026-01-21T15:32:11","slug":"an-ai-run-company-just-revealed-what-our-work-future-might-really-look-like","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/us\/421032\/","title":{"rendered":"An AI-run company just revealed what our work future might really look like"},"content":{"rendered":"

A team of researchers built a fake company staffed entirely with artificial intelligence agents to see whether machines could really replace human employees. The findings should reassure anyone worried about losing their job to AI.<\/p>\n

Is artificial intelligence<\/a> on the verge of taking over our jobs? Some companies are betting on it, while others remain skeptical, calling the technology overhyped. So, what\u2019s the truth? In a preprint study posted on Arxiv, researchers from Carnegie Mellon University created a simulated company and hired AI agents to run it. The outcome, however, was anything but encouraging.<\/p>\n

The \u201cstaff\u201d included agents based on Anthropic\u2019s Claude, OpenAI\u2019s GPT-4o, Google\u2019s Gemini, Amazon\u2019s Nova, Meta\u2019s Llama, and Alibaba\u2019s Qwen. Each was assigned a role such as financial analyst, project manager, or software engineer. Meanwhile, the researchers used another platform to simulate coworkers the agents had to contact for certain tasks \u2014 including an HR department.<\/p>\n

The agents failed more than three-quarters of their assignments<\/p>\n

The agents were given a range of jobs, from analyzing databases to conducting virtual tours of office spaces. Claude 3.5 Sonnet was the top performer, but even it managed to complete only 24 percent of its tasks. Counting partial completions, that number rose to just 34.4 percent. Gemini 2.0 Flash came next, finishing 11.4 percent \u2014 and no other agent exceeded 10 percent.<\/p>\n

In terms of costs, Claude 3.5 Sonnet racked up $6.34 in operational expenses, compared to Gemini 2.0 Flash\u2019s modest $0.79. Still, the overall results make it clear that today\u2019s AI tools \u2014 for all their speed and efficiency \u2014 are far from capable of truly autonomous work.<\/p>\n

The researchers noted that many failures came from misunderstanding subtle instructions. When told to save a file with a \u201c.docx\u201d extension, for instance, the AIs didn\u2019t recognize it as a Microsoft Word document. They also stumbled on tasks requiring communication<\/a> or social reasoning. The biggest challenge came when browsing the web \u2014 especially handling pop-ups. And when the systems got lost, they often took shortcuts, skipping difficult steps and mistakenly assuming success.<\/p>\n

These findings suggest that while artificial intelligence can excel at specific and narrowly defined tasks, it\u2019s still a long way from working independently. For now, human judgment, creativity, and adaptability remain vital parts of any workplace.<\/p>\n

\"\"<\/p>\n

Edward Back<\/p>\n

Journalist<\/p>\n

My passion for programming began with my very first computer, an Amstrad CPC 6128. I started coding in Basic, then moved on to Turbo Pascal on a 286, eventually exploring more modern languages including web development. I\u2019m also deeply interested in science, which led me to attend a math-focused preparatory program. Later, I studied psychology with a focus on the cognitive aspects of artificial intelligence.<\/p>\n","protected":false},"excerpt":{"rendered":"A team of researchers built a fake company staffed entirely with artificial intelligence agents to see whether machines…\n","protected":false},"author":2,"featured_media":421033,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[45],"tags":[182,181,507,74],"class_list":["post-421032","post","type-post","status-publish","format-standard","has-post-thumbnail","category-artificial-intelligence","tag-ai","tag-artificial-intelligence","tag-artificialintelligence","tag-technology"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/posts\/421032","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/comments?post=421032"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/posts\/421032\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/media\/421033"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/media?parent=421032"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/categories?post=421032"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/tags?post=421032"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}