This week in AI: telling real from generated
Google has expanded its SynthID Detector worldwide and the EU's transparency rules for AI-generated content are coming into effect. Public bodies should be clear about where they use AI.
Read articleA study of a medical AI in primary care was published in The Lancet, labs reported progress on drug development, and the UK commissioned a review of NHS productivity and AI. Health is where AI evidence is maturing fastest.
In October, Google published research in The Lancet on AMIE, a conversational medical AI, used during primary care visits. Anthropic reported early progress on using Claude to speed up drug development. And the Budget confirmed an independent review of NHS productivity, including the role of AI, due to report in early 2027.
Health care is where AI evidence is maturing fastest, because the sector insists on it. A study in a peer-reviewed journal, with a clear method and published results, is worth more than any number of demos.
That standard is worth borrowing elsewhere. When a council or housing association considers an AI tool, it is fair to ask: what was it tested on, how was success measured, and what went wrong? A good supplier will have answers.
In the NHS itself, we think the quickest gains are likely to come from administration rather than diagnosis: letters, referrals, scheduling and coding. These tasks take up a large share of staff time, they are well defined, and errors can be caught before they reach a patient. Freeing up time there helps clinicians without asking AI to make clinical decisions.
We look forward to the productivity review's findings next year.
Keep reading
Google has expanded its SynthID Detector worldwide and the EU's transparency rules for AI-generated content are coming into effect. Public bodies should be clear about where they use AI.
Read articleThe Autumn Budget on 28 October included a package of AI measures, from more funding for the AI Security Institute to local AI partnerships with mayoral authorities. Here are the parts most relevant to public services.
Read articleAnthropic has donated Agent Skills to the Agentic AI Foundation, joining the Model Context Protocol. Shared standards make it easier to switch tools and keep what you have built.
Read article