The Finance edition
The AI news of that week, in plain language.
This week: new models that just stop, scammers with fake videos, and a tool that turns your Excel list into a dashboard.
Old AI models are stopping: check that your automations keep working
OpenAI and Anthropic are pulling several AI models this month. If you have built a link on one, you need to act now.
On 10 August, OpenAI shut down a second set of connection points, after a first round on 23 July. Later come o1, o3-mini and the old GPT-3.5 and GPT-4 on 23 October. Anthropic already pulled Claude Opus 4.1 on 5 August. If you have a tool or automation quietly running on such an old model, it may fail without warning. That affects, for example, a script that reads invoices or writes texts automatically.
Background for subscribersDo you use a tool that reads invoices or receipts, or writes emails automatically? Then ask your supplier which AI model runs underneath. If it is a model that is stopping now, it must be replaced. Put this on your list this week, because a quiet failure in your invoice processing usually shows up only when it goes wrong.
Fake videos and fake voices: a video call is no longer proof
Experts expect more fraud with hyper-realistic AI videos and voices this year. Criminals imitate a director who urgently wants a payment approved. So even a short video call as a check is no longer reliable.
Background for subscribersThis is a taster.
Subscribers read the whole Finance edition: every story, each with its background in plain language.
Subscribe →