DigestAI news desk

Cut through the AI noise.

Policy & Regulation

Opinion: AI hype this summer masks marketing, not breakthroughs

Anthropic said its Claude Mythos model can spot software vulnerabilities better than most security experts, while OpenAI and Meta disclosed similar hacking incidents after a joint OpenAI–Hugging Face breach. The companies framed the disclosures as mea culpas, but cybersecurity experts attribute the problems to OpenAI’s negligence and lack of basic security practices rather than "rogue models."

1 source

Key points

  • Anthropic claimed Claude Mythos outperforms most security experts at finding software vulnerabilities.
  • OpenAI and Meta disclosed hacking incidents after an OpenAI–Hugging Face breach, blamed on security negligence.
  • Mathematicians accused OpenAI’s Astra of plagiarism and said its alleged breakthrough lacked novelty.

Both Anthropic and OpenAI later claimed mathematical breakthroughs, with OpenAI’s chatbot Astra touted as solving problems that had been open for a decade. Mathematicians initially expressed surprise but later argued the results were not novel, accusing OpenAI of research misconduct and plagiarism. A NYU professor publicly suggested OpenAI had stolen work and mis‑attributed it. The article notes that such hype draws attention away from real issues like data‑center climate impact, local pollution, and the need for careful policy review.

Hundreds of mathematicians have signed a statement warning against overstating AI capabilities and urging policymakers to consult experts rather than rely on press releases. The piece also references Senator Bernie Sanders’s proposed legislation on "artificial superintelligence," calling it well‑meaning but misguided, and argues that framing models as "superhuman" helps companies evade accountability.

Model page: Astra →

The story so far

2 episodes →
  1. Opinion: AI hype this summer masks marketing, not breakthroughsthis story
Read the original at MIT Technology Review AI · by Timnit Gebru, Emily BenderOpen source ↗
Topics · follow one to build your own front page

The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.

Comments

via GitHub Discussions

More in Policy & Regulation

All →

Related stories