OpenAI Ends the Honor System for User Age Verification

Because OpenAI does not want to rely on kids being honest about their age anymore. ChatGPT is rolling out automated age prediction to identify users who are likely under 18 so it can apply stricter safety rules by default. Instead of self reporting, the system looks at signals like account age, usage patterns, and when someone is active to decide if extra protections are needed. If a user is flagged as a minor, ChatGPT automatically blocks things like graphic violence, sexual roleplay, and content tied to unhealthy body standards.

It means safety controls will be applied more aggressively, even if someone says they are an adult. Adults who are incorrectly flagged can appeal by verifying their age through a selfie or government ID using a third party service called Persona, which deletes biometric data after verification. This move also helps OpenAI respond to regulator pressure and sets the stage for a future “adult mode” that separates mature content from younger users. The rollout started January 19, with EU users seeing it later due to privacy rules, and it raises bigger questions about how much AI can infer about people based on behavior alone.

DeepSeek is getting ready to release its next big AI model, about a year after its R1 launch flipped the industry on its head. That earlier model shocked investors by showing that strong AI did not need massive, expensive compute clusters. The community is anticipating a mid February release of DeepSeek V4, with its new ‘Model1’ architecture.

The upcoming model is expected to be built around a new architecture designed for long context coding. Early info suggests it can handle over one million tokens, which would let developers work with entire codebases in one prompt. That kind of scale could change how coding assistants are used day to day. If DeepSeek pulls it off again, it reinforces the idea that lean teams can outpace bigger, more expensive labs.

🚢 Experts debate the impact of exporting Nvidia chips to China

☢️ Anthropic CEO compares chip exports to nuclear proliferation

🗣️ Apple’s new Siri could compete with ChatGPT and Claude

⚠️ JPMorgan CEO warns AI could trigger civil unrest

🛑 Major charities stop using X over AI safety concerns

🥑 Meta Superintelligence Labs delivers its first internal models

Challenge: Use AI to gain some perspective on your life by interviewing your future self. MIT research suggests that speaking with a simulated version of yourself can significantly reduce anxiety and help you make better long term decisions.

Steps:

1️⃣ Visit the Future You platform or use your favorite chatbot. If you are using a chatbot, provide it with a detailed "memory" of your current values, goals, and daily struggles.

2️⃣ Ask the AI to adopt the persona of your 60 year old self. Instruct it to answer based on a life where you successfully navigated your current challenges.

3️⃣ Ask three specific questions: "What is one thing I am worrying about now that won't matter in ten years?", "What habit should I start today to make your life easier?", and "What is the most important lesson you learned between our ages?"

Spend at least 15 minutes in this dialogue. The goal is to bridge the gap between who you are now and who you want to become.

Will automated age checks make AI safer for kids, or is this just more data collection? We want to hear your thoughts on the new ChatGPT safeguards!