Daily AI Newsletter: Sam Altman says ChatGPT will soon allow erotica for adult users

Loading page…

In today’s email:

👨🏻‍🏫 Inside San Francisco’s new AI school: is this the future of US education?
🩺 Should an AI copy of you help decide if you live or die?
🕳️ OpenAI hires black hole physicist in broader science push
🧰 10 new AI-powered tools and resources. Make sure to check the online version for the full list of tools.

Sam Altman says ChatGPT will soon allow erotica for adult users

Key Takeaway: OpenAI plans to relax ChatGPT’s safety rules—letting verified adults engage in erotic chats and tweak the bot’s tone—while critics warn the shift could endanger vulnerable users.

More Insights:

Altman announced the change on X (Oct 14, 2025), tying it to expanded age-gating and “treat adult users like adults.”
Adult access will rely on an age-prediction system, with possible ID upload if misclassified—raising privacy trade-offs.
OpenAI cites progress on “sycophancy” risks: GPT-5 (launched in August) and new safety tools, plus a mental-health expert council.
Controversies linger: earlier GPT-4o incidents and a lawsuit over a teen’s suicide fuel doubts that risks are “under control.”
Market pressure looms: 800M weekly active users and competition with Google/Meta; erotic role-play has boosted engagement at rivals like Character AI.

Why it matters: This sets a precedent for mainstream AI to allow sexual content—testing whether platforms can expand adult freedoms without amplifying harms to minors and at-risk users, and forcing a hard reckoning over identity verification, privacy, and growth-vs-safety trade-offs.

100+ ChatGPT Prompts to Revolutionize Your Day

Supercharge your productivity with HubSpot's comprehensive guide. This free resource is your fast track to AI mastery:

Industry-Specific Use Cases: 15+ real-world applications across various sectors
Productivity Guide: 21 best practices to 10x your efficiency with AI
Prompt Powerhouse: 100+ ready-to-use prompts for immediate implementation
Challenge Buster: Overcome common AI hurdles with expert strategies

Plus, in-depth sections on email composition, content creation, customer support, and data analysis.

Find Your Guide Here

OpenAI has five years to turn $13 billion into $1 trillion

Key Takeaway: OpenAI has a five-year plan to finance more than $1 trillion in AI infrastructure commitments through new revenue lines (enterprise/government deals, Sora, agents, shopping, ads, and hardware), “creative” debt, and deep chip partnerships—betting that scale and falling costs will close the gap.

More Insights:

Revenue today: ~$13B ARR; ~800M users with ~5% paid—aiming to double paid share; rolling out lower pricing in India and expanding to the Philippines, Brazil.
New lines: bespoke government/enterprise products, checkout-driven shopping, cautious advertising, Sora video and AI agents, plus a Jony Ive–designed consumer device.
Infra strategy: staggered capacity deals with Nvidia/AMD/Oracle/Broadcom (~26GW over a decade), potential “Stargate” compute supply role, and IP-driven infra plays.
Financing: leveraging partners’ balance sheets now, exploring novel debt structures and chip-financing; critics note circularity, but signals are “we’re good for the debt.”
Risk/return math: ~20GW ≈ ~20 nuclear reactors’ power; H1 operating loss ~$8B even as revenue more than doubled; leadership thinks big compute could nearly scale revenue.

Why it matters: If this works, OpenAI could reset the economics of chips, power, and software distribution—blurring the line between AI product, ad platform, and utility; if it doesn’t, the unwind could jolt tech capex, valuations, and energy planning across the economy.

Inside San Francisco’s new AI school: is this the future of US education?

Key Takeaway: Alpha School promises kids can learn twice as fast with just two hours of daily academics powered by AI, but experts say the evidence is thin, the benefits uneven, and equity concerns loom.

More Insights:

AI mainly personalizes pacing and assignments; it’s guide-led, not chatbot-taught.
Model echoes older self-directed approaches (e.g., Montessori) and long-used tools (IXL, Khan, Duolingo, Math Academy).
Claims of top 1–2% scores and 90% satisfaction face selection-bias questions given affluent demographics and sky-high SF tuition.
Researchers urge rigorous trials and warn about hallucinations, bias, and risks for less-motivated or younger learners.
Public districts are cautiously integrating AI literacy and pilots, signaling inevitability—but not a one-size-fits-all solution.

Why it matters: If elite, AI-heavy models set the narrative without solid evidence and guardrails, schools could hard-code inequality into the next era of learning—making careful evaluation and equitable access the real test of “AI in education.”