📌 Apple is reportedly developing a wearable AI pin

📜 Anthropic publishes Claude’s Constitution

🎧 OpenAI hints at its first AI hardware device

🏥 Amazon launches AI health-care tool

🎵 ElevenLabs releases full AI co-created music album

+ 📊 Daily poll and results

+ 📈 Trending tools and resources

Join the Superhuman AI newsletter - read by 1M+ professionals

Learn AI skills in 3 mins a day

Become the AI expert on your team

Apple is reportedly working on a wearable AI pin that looks like a thicker AirTag and includes cameras, microphones, a speaker, and wireless charging, according to reports from The Information and Bloomberg, though the project is still early and could be canceled.

The reports say Apple is also planning a major Siri overhaul into a ChatGPT-style chatbot, and with Google’s Gemini set to power Siri AI, the company appears to be making a more serious push into generative AI.

The AI pin could launch as early as 2027 with plans to produce around 20 million units, but Apple may face challenges around privacy concerns and convincing users after the high-profile failure of Humane’s AI pin.

Anthropic published Claude’s Constitution, a document written directly to the AI that sets clear priorities in this order: be safe, act ethically, follow company rules, and then help users.

Instead of strict do’s and don’ts, the document explains the reasoning behind each principle so Claude can apply these values to new or unexpected situations.

The constitution also talks about Claude’s well-being and even suggests the AI could matter morally, including a rare rule telling Claude to refuse or disobey Anthropic if asked to do something unethical.

OpenAI says it plans to announce its first hardware product in the second half of this year, following last year’s acquisition of Jony Ive’s startup, with executives describing the device as calmer and more peaceful than a smartphone.

Reports and leaks suggest the product could be a pair of screen free, pocketable AI earbuds codenamed “Sweet Pea,” possibly running on a custom 2 nanometer chip to handle AI tasks locally instead of relying on the cloud.

With ChatGPT nearing a billion weekly users, OpenAI wants more control over how its AI is distributed, but breaking into daily use will be tough given past AI device failures and strong competition from existing earbuds and new wearable tech.

Amazon rolled out Health AI, a new assistant for One Medical members that uses AI models from Amazon Bedrock to answer questions, manage medications, book appointments, and give personalized guidance based on medical records, labs, and prescriptions.

The tool lives inside the One Medical app, does not add conversations to official medical records, and is designed to escalate users to a provider or in-person visit when symptoms require it, rather than replacing doctors.

Health AI was tested with a small group last year and is now expanding more broadly, as Amazon deepens its push into health care following its One Medical and PillPack acquisitions and amid growing competition from other AI health tools.

ElevenLabs released a 13-track album made with its Eleven Music model, co-created with human artists like Liza Minnelli and Simon Garfunkel as a proof-of-concept for ethical human and AI collaboration.

The project spans rap, EDM, spoken word, and Brazilian funk, with artists keeping full ownership and all streaming royalties, whether tracks are fully AI-made or combine AI tools with human performances.

The release highlights how AI music is shifting from lawsuits and backlash to industry acceptance, as major labels now partner with AI companies and artists increasingly use the tech as a creative tool rather than a threat.

A) Yes, short clips are highly addictive - 62% 🏆

B) No, users will still come mainly for full shows - 38%