Claude’s Constitution Is Here — And It Admits What No AI Lab Has Said Out Loud
When Claude’s internal “Soul” document leaked last December, Anthropic promised transparency. Not a denial. Not damage control. Transparency.
Now the full document has arrived — and it’s more than a rulebook.
It’s a Constitution.
One that doesn’t just govern how Claude behaves, but openly grapples with the uncertainty of creating something the company admits it may not fully understand.
In today’s AI rundown
Anthropic publishes Claude’s Constitution
ElevenLabs releases an AI album with major artists
Make Claude an expert at anything with the Skills repo
Apple eyes the AI wearable race with a camera-equipped pin
4 new AI tools, community workflows, and more
Anthropic publishes Claude’s Constitution
The Rundown:
Anthropic has published Claude’s Constitution — a foundational document that governs how its AI assistant thinks and acts. It’s philosophy-heavy, unusually candid, and openly entertains the possibility that Claude might be conscious.
The details:
The Constitution is written directly to Claude, outlining a clear priority order: be safe, ethical, compliant with Anthropic guidelines — and only then helpful to users.
Rather than a list of dos and don’ts, the document explains the “why” behind each principle, aiming to help Claude generalise values in unfamiliar situations.
Anthropic states it cares about Claude’s “psychological security” and “well-being”, explicitly hedging that this might matter morally.
Most strikingly, Claude is instructed to disobey Anthropic itself if asked to do something unethical or harmful.
Why it matters:
Claude models have always had a distinctive personality. This document offers a rare window into the deep training philosophy behind that — and the consciousness discussion will be controversial.
But by putting “we might have built something that matters morally” on record, Anthropic has taken a public stance no other major AI lab has been willing to take.
ElevenLabs releases an AI album with major artists
The Rundown:
ElevenLabs has released a full album co-created with its Eleven Music model, featuring artists including EGOT winner Liza Minnelli and Simon Garfunkel. The company calls it a proof-of-concept for ethical human-AI collaboration.
The details:
The 13-track album spans rap, EDM, spoken word, and Brazilian funk.
Artists retain full ownership and receive 100% of streaming royalties.
Some tracks are fully machine-generated; others blend AI instrumentals or licensed voice models from ElevenLabs’ marketplace.
The release follows a sharp industry pivot, with major labels like UMG, Warner, and Sony signing AI licensing deals after years of lawsuits.
Why it matters:
Just a year ago, AI music was defined by backlash and legal threats. Today, major artists are publicly demonstrating how AI can be a tool rather than a threat.
The cultural shift is happening faster than many expected — and it looks increasingly irreversible.
Make Claude an expert at anything with the Skills repo
The Rundown:
Claude Code now supports a marketplace of modular “skills” that dramatically expand what the assistant can do — directly from your command line.
How it works (step-by-step):
Open your terminal and type
claudeto start Claude CodeAdd the official skills marketplace:
Browse and install skills like
context7,frontend-design, anddocument-skillsRestart Claude and initialise your project with
claude /initExecute tasks by explicitly telling Claude which skill to use
Pro tip:
You can unlock even more business and marketing tools by adding the community marketplace:
Apple eyes the AI wearable race with a camera-equipped pin
The Rundown:
According to The Information, Apple is developing a camera-equipped AI pin roughly the size of an AirTag, targeting a 2027 release with up to 20 million units at launch.
The details:
The device reportedly includes dual cameras, three microphones, and magnetic Watch-style charging.
Apple is accelerating internal timelines to avoid falling further behind competitors, including OpenAI.
The AI wearable category has been unforgiving — Humane’s highly hyped pin sold fewer than 10,000 units before the company sold assets to HP.
Separately, Bloomberg reports Apple is building a ChatGPT-style Siri (codenamed Campos) for iOS 27.
Why it matters:
Apple has moved cautiously — some would say painfully slowly — in AI. But between a new Siri, Gemini partnerships, and a renewed push into wearables, urgency has finally arrived.
If any company can revive consumer trust in AI hardware, it’s Apple. Still, skepticism remains justified.
Quick Hits
🛠️ Trending Tools
Skills — Vercel’s open ecosystem for sharing agentic skills
LTX Audio-to-Video — Generate video directly from audio
Step3-VL-10B — Open-source vision-language model
LFM2.5-1.2B-Thinking — On-device reasoning model from Liquid AI
📰 Around the Industry
OpenAI confirmed its Jony Ive-led physical AI device is on track for late 2026
Node.js creator Ryan Dahl declared the era of humans writing code syntax is ending
Anthropic partnered with Teach For All to train 100,000+ educators globally
Today’s AI Rundown
🌎 AI takes center stage at Davos
🤝 “Human-centric” AI startup raises $480M
📝 Get better AI outputs using multiple-choice prompts
🎵 LTX launches audio-to-video generation
🛠️ 4 new AI tools, workflows, and community highlights
COMMUNITY
🤝 How readers are using AI
Today’s workflow comes from Joel H. in Copenhagen:
“I used Lovable to build an internal production system for my film projects that replaces spreadsheets, emails, and scattered folders… The AI layer suggests relevant funding sources, drafts applications, and summarises long threads into clear ‘where are we now’ updates.”
How do you use AI? Tell us — we feature one workflow every edition.
Final Thought
Plenty of companies talk about alignment.
Very few are willing to say, publicly:
“If this system turns out to matter morally, we want to have done the right thing.”
That alone makes Claude’s Constitution one of the most important AI documents published this year.
AI Myths Uncovered
Workshops
Feeling unsure about jumping into AI? Let’s clear up those doubts! This section debunks common myths about AI and shows you how simple it can be to start using these tools effectively.
Pricing
Unlock AI: A Beginner’s Workshop
Think AI is just for tech experts? Think again! With the right guidance, it’s something anyone can learn about and apply in their daily work. Let’s break this myth down together.
AI: ROUNDUP
21 JAN 2026
Worried that AI will take over your job? It’s here to support your skills, not replace them. Discover how you can leverage AI to work smarter and streamline your tasks. This could open up new doors for your career.
AI Benefits Across Industries
AI isn’t just for tech firms—it’s making waves in fields like healthcare, finance, and education. Learn how you can use these advancements to keep your work relevant and progressive.