The Download: meet our AI innovators, and what happens when therapists use AI covertly

Summary

Today’s edition of The Download highlights MIT Technology Review’s 35 Innovators Under 35 for AI in 2025, profiles Yichao “Peak” Ji and the viral Manus app, and examines a worrying trend: therapists secretly using ChatGPT during sessions. The newsletter also flags upcoming events, must-read stories across tech and geopolitics, and a longer piece on how generative AI is reshaping internet search.

Source

Source: https://www.technologyreview.com/2025/09/09/1123447/the-download-meet-our-ai-innovators-and-what-happens-when-therapists-use-ai-covertly/

Key Points

MIT Technology Review named its 2025 AI honourees in the 35 Innovators Under 35 list — people shaping model development and novel AI techniques.
Yichao “Peak” Ji, chief scientist at Butterfly Effect, helped make the Manus AI agent a viral hit, attracting millions to its waitlist.
Several therapists have been found to use ChatGPT covertly during sessions; that raises ethical, safety and privacy concerns because consumer models aren’t clinically vetted.
The newsletter highlights a LinkedIn Live event on tech breakthroughs (quantum, humanoid robots, AI agents, climate tech) on 10 September.
Top news round-up includes US retreat from disinformation collaboration with Europe, legal scrutiny around Anthropic, a whistleblower lawsuit at Meta/WhatsApp, and debates over AI regulation and the future of search.

Content summary

The Download aggregates several short pieces and links. It showcases the AI innovators list and links to full profiles, including a feature on Yichao Ji and the rapid rise of Manus. The newsletter draws attention to a report about therapists secretly consulting ChatGPT during live sessions — sometimes awkwardly exposed when clinicians accidentally shared screens. Reporters note that while AI could play a legitimate therapeutic role, covert use of unvetted models creates risks for patients.

Beyond those headlines, the edition promotes a live editorial discussion on transformative technologies and curates ten must-read items covering disinformation policy, legal battles in AI, whistleblower claims at Meta, government tech appointments, and broader questions about the evolution of search in the age of generative models.

Context and relevance

This newsletter is useful if you follow AI’s social and regulatory impacts. The innovators list signals where talent and research are concentrating; the Manus profile shows how product storytelling and usability can create global momentum for AI apps. The therapist-ChatGPT story is a practical warning: conversational models are being used in sensitive settings without clinical oversight, which intersects with privacy, professional ethics and liability debates. The generative-search piece frames a longer-term change in how online information will be delivered and monetised.

Why should I read this?

Short version: we read the scattershot stuff so you don’t have to. If you care about where AI talent is heading, what breakout products look like, and the messy real-world ways people are already using (and misusing) these models — this edition gives you the headlines and the links to dig deeper. Also, if you’re worried about ethics in mental-health care or the future of search, the therapist/ChatGPT and generative-search items are worth a quick click.

Author

Punchy — we’ve pulled the key bits and signposted the full reads so you can pick what to follow up on. If something’s crucial, we point it out; if it’s background noise, we saved you the scroll.