Study: ChatGPT beats other AIs at creating fake news — and what business should do about it

ER10 relayed a study in which scientists compared popular AI models' ability to generate disinformation: ChatGPT wrote the most convincing fake news of all. The irony is that the very quality the model is valued for at work — coherent, confident, "journalistic" text — also makes it the best instrument of deception.

For society this is a regulation question, and AI-content labeling in the spirit of the European AI Act is an attempted answer. But the problem has a corporate dimension people consider less often: a fake story about a company — a "report" of toxic production, a "leak" of client data, a "bankruptcy" — is now written in a minute and spreads faster than any rebuttal. The target may be not a ministry but a regional developer or a pharmacy chain — for instance, at the hands of an unscrupulous competitor.

Practical preparation for a business in Kazakhstan requires no crisis agency budget. First: set up monitoring of company mentions — search and social media alerts on your name, so you learn about a fake within hours, not days. Second: choose in advance the channel your customers trust — the official website, a mailing list, a messenger — and keep it alive in calm times. A rebuttal from a channel that has been silent for years looks suspicious in itself.