The Finance edition
The AI news of that week, in plain language.
This week: fake invoices that look genuine, new rules for AI, and three ChatGPT models you can use smartly to keep costs down.
Fake payment requests with AI: one phone call prevents a costly mistake
Fraudsters use AI to create convincing fake videos and voice clones to force payments.
Criminals are increasingly using deepfake videos and cloned voices for CEO fraud. An employee transfers money, believing they are speaking to their actual boss. Experts say that standard security measures (such as logging in with a code) offer no protection here. The remedy is procedure, not technology: dual approval for payments and verification through a second channel.
Background for subscribersMake one firm rule with your team: you never approve an urgent payment or a changed account number based solely on an email, PDF, voice, or video. If in doubt, call back on the known number, not the number on the invoice. You could also agree on a code word for urgent payment requests. One phone call can prevent a mistake that costs you thousands of euros.
EU AI Act applies from 2 August: chatbots and AI content must be clearly identifiable
From 2 August 2026, the transparency rules of the European AI Act are in force. If you have a chatbot on your website, AI-generated marketing copy, or AI in client communications, you must state this clearly. New AI systems brought to market after this date must comply immediately; older systems have until 2 December 2026.
Background for subscribersThis is a taster.
Subscribers read the whole Finance edition: every story, each with its background in plain language.
Subscribe →