"A host of new AI models the technology that underpins tools like chatbots were released last year, including OpenAI's GPT-5, Anthropic's Claude Opus 4.5 and Google's Gemini 3. The report points to new reasoning systems which solve problems by breaking them down into smaller steps showing improved performance in maths, coding and science. Bengio said there has been a very significant jump in AI reasoning. Last year, systems developed by Google and OpenAI achieved a gold-level performance in the International Mathematical Olympiad a first for AI."
"However, the report says AI capabilities remain jagged, referring to systems displaying astonishing prowess in some areas but not in others. While advanced AI systems are impressive at maths, science, coding and creating images, they remain prone to making false statements, or hallucinations, and cannot carry out lengthy projects autonomously. Nonetheless, the report cites a study showing that AI systems are rapidly improving their ability to carry out certain software engineering tasks with their duration doubling every seven months."
"If that rate of progress continues, AI systems could complete tasks lasting several hours by 2027 and several days by 2030. This is the scenario under which AI becomes a real threat to jobs."
"The report describes the growth of deepfake pornography as a particular concern, citing a study showing that 15% of UK adults have seen such images. It adds that since the publication of the inaugural safety report in January 2025, AI-generated content has become harder to distinguish from real content and points to a study last year in which 77% of participants misidentified text generated by ChatGPT as being human-written."
Multiple advanced AI models were released last year, including GPT-5, Claude Opus 4.5 and Gemini 3. New reasoning systems that decompose problems into smaller steps improved performance in mathematics, coding and science and reached gold-level performance at the International Mathematical Olympiad. Capabilities remain uneven: systems show exceptional skill in specific tasks but still produce false statements and hallucinations and cannot autonomously run lengthy projects. Software engineering abilities are improving rapidly, with task-duration capability doubling every seven months, which could enable hour- and multi-day task automation by 2027–2030 and raise job-displacement concerns. Deepfake pornography and harder-to-detect AI-generated content are growing problems.
Read at www.theguardian.com
Unable to calculate read time
Collection
[
|
...
]