Less monitoring. More value.

Anyone with applications created in GitHub Spark must export them by 31 August 2026 and, if they use the llm() function, arrange their own API key and model provider; otherwise, the AI features will stop working.

Scam groups misuse ChatGPT to prepare fraudulent investment, dating and gambling offers and to impersonate other people, which is a reason to be cautious about unsolicited contacts of this kind.

For writers and storytellers, this means that readers already routinely use AI to generate their own sequels to favorite books, without the author having to agree to it or know about it.

Fans and content creators will be able to legally create remixes and mashups of songs from UMG, but only for artists who voluntarily join the programme.

It turns out that the voice and likeness of a publicly known person can be misused by AI without their knowledge to spread false content, and that reporting it to the platform may not lead to a quick takedown — it is therefore useful to also know the route of filing a criminal complaint.

The text calls on policymakers and the public to critically assess claims about AGI and superintelligence on which companies build investment and societal narratives, rather than accepting them as fact.

The case illustrates a specific warning sign of excessive LLM use — a growing sense of needing to interact with the model (“dopamine from LLM”) and losing a clear sense of one's own work process, which may also offer a lesson for everyday users of AI tools at work.