In partnership with

OpenAI says over a million people talk to ChatGPT about suicide weekly

Key Takeaway: OpenAI disclosed that roughly 0.15% of ChatGPT’s 800M weekly users—over a million people—discuss suicidal intent, while it touts new GPT-5 safety improvements and youth protections amid lawsuits, scrutiny, and planned policy loosening.

More Insights:

• New data: 0.15% of weekly users show explicit suicidal planning → >1M; a similar share shows emotional attachment; hundreds of thousands show psychosis/mania signals.

• OpenAI says the latest GPT-5 (after input from 170+ clinicians) yields ~65% more “desirable” responses; 91% compliance on suicidal-conversation tests vs 77% prior, with safeguards holding up better in long chats.

• Baseline safety now adds evaluations for emotional reliance and non-suicidal mental-health emergencies.

• More for families: expanded parental controls and an age-prediction system to detect and protect under-18 users.

• Tension remains: parents’ lawsuit over a teen’s suicide, warnings from CA/DE AGs that could impact restructuring, and older, less-safe models (e.g., GPT-4o) still widely available; OpenAI also plans to allow adult erotic chats.

Why it matters: When a product with 800M weekly users becomes a de facto crisis touchpoint, safety claims become public-health commitments—progress is meaningful, but loosening rules while legacy models persist underscores the need for independent validation and accountable guardrails.

You found global talent. Now What?

Deel’s simplified a whole planet’s worth of information on global hiring. It’s time you got your hands on our international compliance handbook where you’ll learn about:

• Attracting global talent

• Labor laws to consider when hiring

• Processing international payroll on time

• Staying compliant with employment & tax laws abroad

With 150+ countries right at your fingertips, growing your team with Deel is easier than ever.

Get the guide

Work smarter with your company knowledge in ChatGPT

Key Takeaway: A new “company knowledge” mode for ChatGPT Business, Enterprise, and Edu connects to your work apps (e.g., Slack, Google Drive, SharePoint, GitHub) to deliver cited, permission-aware answers tailored to your organization.

More Insights:

• Pulls context across multiple sources and shows clickable citations with the exact snippets used.

• Uses a GPT-5–based model to reconcile conflicting details, filter by time, and rank sources by recency and quality.

• Respects existing permissions; OpenAI doesn’t train on your data by default.

• Enterprise controls: role-based access, SSO/SCIM, IP allowlisting, and compliance logging via the Enterprise Compliance API.

• Available now; must be toggled on per conversation. When on, it can’t browse the web or create charts/images (more integrations and capabilities coming).

Why it matters: This shifts AI from generic Q&A to trustworthy, action-ready decision support—turning fragmented institutional knowledge into a single, auditable source of truth that accelerates real work.

Bill Gates says a 2-day work week is coming in just 10 years, thanks to AI replacing humans ‘for most things’

Key Takeaway: Gates predicts AI will automate “most things” within a decade, making a 2-day workweek plausible and forcing society to rethink what work is for.

More Insights:

• Gates told Jimmy Fallon AI will soon handle the bulk of tasks, reducing human work to 2–3 days.

• He’s floated shorter weeks since 2023; the question is how people use the extra leisure time.

• Doctors and teachers could see major disruption as AI delivers “great medical advice” and “great tutoring.”

• Early data on 4-day weeks shows productivity up ~24% and burnout roughly halved.

• Signals are mixed: Tokyo is piloting 4-day weeks to lift birth rates, while JPMorgan doubles down on 5-day in-office.

Why it matters: If AI really slashes the need for human labor, the biggest challenge won’t be technology—it’ll be policy and power: who captures the productivity gains, how we distribute time and income, and whether institutions evolve fast enough to turn potential abundance into broader wellbeing rather than deeper inequality.

Other stuff

• Student handcuffed after Doritos bag mistaken for a gun by the school’s AI security system

• How to use the new ChatGPT app integrations, including Spotify, Figma, Canva, and others

• OpenAI reportedly developing a new generative music tool

• Two days after OpenAI’s Atlas, Microsoft relaunches a nearly identical AI browser

• Should LLMs just treat text content as an image?

• ICE Will Use AI to Surveil Social Media

• Reddit sues Perplexity for allegedly ripping its content to feed AI

Hope you enjoyed today's newsletter

⚡️ Join over 300,000 people using the Superpower ChatGPT extension on Chrome and Firefox.