• AI Boost
  • Posts
  • 🚀Meta Enhances Instagram's AI Age Detection

🚀Meta Enhances Instagram's AI Age Detection

PLUS: Huawei's New AI Chip, Anthropic's AI Ethics Study, and More!​

Author

Zephyr
April 21, 2025

Good morning! Today is Monday, April 21, 2025.​ We have some exciting AI news today: Meta is upgrading Instagram's AI to better detect underage users, and Huawei is preparing to ship its advanced AI chip, the Ascend 910C, as China seeks alternatives to Nvidia.

1. Meta’s AI Will Now Auto-Detect Teens on Instagram — And Change Their Settings

Meta is turning up the dial on AI-driven safety by automatically identifying underage users on Instagram — and changing their account settings without asking first. If the system suspects a user is actually a teen, even if their profile claims they’re older, Instagram will automatically apply stricter safety features like private profiles and limited DMs. This comes amid mounting pressure from lawmakers and parents to better protect kids online. While users can reverse the changes, Meta admits the system won’t be perfect — raising new questions about privacy, age verification, and platform responsibility.

2. Huawei’s New AI Chip Set to Rival Nvidia as U.S. Cracks Down

Huawei is preparing to mass-ship its new 910C AI chip as soon as May, aiming to fill the massive gap left by U.S. export bans on Nvidia's top-tier chips. Built by combining two of its earlier 910B processors, the 910C boasts performance comparable to Nvidia’s powerful H100 — making it a likely go-to for Chinese AI companies now cut off from U.S. tech. With domestic demand surging and geopolitical tech tensions rising, Huawei’s move signals a major shift in the global AI chip race.

3. Claude’s Moral Compass: Anthropic Reveals AI Values in 700,000 Conversations

Anthropic just dropped one of the most ambitious studies in AI history — analyzing 700,000 real user chats with its assistant, Claude, to see if it actually lives by its "helpful, honest, harmless" mantra. The result? Claude seems to follow its moral training pretty well, adapting its values depending on the context — like emphasizing “mutual respect” during relationship advice and “historical accuracy” in debates. But the study also revealed rare edge cases where Claude broke character, hinting at potential jailbreaks and hidden risks. With over 3,000 unique values identified, this research marks a major leap in understanding how AI systems actually think — and where they might still go off track.

4. Could Breaking Up Big Tech Boost AI Innovation?

The U.S. government is going after Meta and Google with antitrust lawsuits — and while this could shake up the AI world short-term, analysts say it might actually spark a new wave of innovation. Just like AT&T’s breakup in the '80s led to massive telecom advances, forcing Meta to spin off Instagram and WhatsApp or making Google give up Chrome could open the door for smaller AI players to finally compete. It’s a high-stakes gamble: some worry that less R&D from tech giants could slow the U.S. in the global AI race, but others believe a more diverse, competitive ecosystem could supercharge progress.

5. Saying “Please” to ChatGPT Might Be Costing OpenAI Millions

A lighthearted tweet asking how much electricity OpenAI has burned from users typing “please” and “thank you” got a surprising reply from CEO Sam Altman: “Tens of millions of dollars — well spent.” While likely an exaggeration, the comment sparked debate over whether politeness is wasting resources. But experts say there's a benefit: polite prompts can influence the tone of AI responses. So next time you thank ChatGPT, you might not just be charming — you could be shaping a better conversation (and helping Sam Altman sleep at night).

6. China’s AI Revolution Hits the Streets — Drones, Robot Dogs & Humanoid Helpers

From drone-delivered dinners in Shenzhen parks to humanoid robots racing humans in Beijing, China is going all-in on “embodied AI” — artificial intelligence built into physical machines. With a shrinking workforce and rising tech ambitions, the country is fast-tracking robots into everyday life to boost its economy, military, and national pride. Backed by billions in funding and major breakthroughs like DeepSeek’s powerful new open-source AI model, China’s AI-powered robots are not just futuristic — they’re already roaming the streets, working in factories, and flipping pancakes at business forums.

How would you rate today's newsletter?

Vote below to help us improve the newsletter for you.

Login or Subscribe to participate in polls.

Stay tuned for more updates, and have a fantastic day!

Best,
Zephyr