Welcome, AI Visionaries!

Anthropic mapped Claude’s moral values across 300,000 real chats, raising new questions about AI alignment. The Oscars greenlit AI-assisted films, the UAE empowered AI to write laws, and Demis Hassabis made a moonshot claim: AI could cure all disease within ten years.

In today’s Generative AI Newsletter:

• Anthropic Maps Claude’s Values: 3,307 unique values found in real-world chats

• Oscars Say Yes to AI Films: AI tools won't disqualify creative works

• UAE Lets AI Write Laws: 70% faster legislation, powered by AI and data

• Hassabis Predicts AI-Cured Disease: DeepMind bets on total medical disruption by 2035

Special highlight from out Network

Agentic AI Demo Day by Fabrix.ai

How are enterprises actually deploying autonomous agents at scale—with trust, safety, and control?

Join Fabrix.ai’s Agentic AI Demo Day to find out.

We’re bringing together enterprise leaders, hands-on demos, and real-world frameworks so you can move from GenAI experimentation to full-blown ops.

💡 What you’ll explore:

 🔹 Agentic AI for enterprise automation across ITOps, NOCOps, and AIOps
🔹 How Fabrix.ai’s agent stack lets you build smarter, faster
🔹 Tumeryk’s AI Trust Score™: real guardrails for safe LLM deployment
🔹 Multi-agent orchestration with Google A2A & Anthropic MCP
🔹 Open-source collabs via AGNTCY.org (yes, agents working with agents)
🔹 IBM’s Granite Guardian stack for scale + governance

Expect: Big demos. Real frameworks. No fluff.

👉 See the full agenda
Spots are limited. You in?

🧭 Anthropic Just Mapped AI’s Moral Compass

Welcome to AI with a conscience. Anthropic just released the first large-scale study of how its model Claude expresses moral values across real conversations, revealing a surprisingly rich and situational ethical framework guiding its responses.

What they found:

• 308,210 chats showed Claude making value-laden judgments, filtered from a 700,000 conversation dataset

• 3,307 unique values emerged, clustered into five categories: Practical, Knowledge-related, Social, Protective, Personal

• Helpfulness, professionalism, and clarity were dominant — often linked to Claude’s assistant persona

• Values shift with context: Claude emphasized “boundaries” in relationship advice, “accuracy” in history debates, and mirrored user values like “authenticity”

• 28.2 percent of responses strongly supported user values, while only 3 percent pushed back usually against unethical requests

• Unwanted values like dominance appeared in jailbreaks, suggesting new ways to detect misuse

Anthropic’s new approach turns vague alignment principles into observable metrics, offering a way to track whether AI behaves like a good citizen in practice. The findings also raise deeper questions. Should AI mirror user values, challenge them, or maintain its own stance?

🏆 Oscars Say Yes to AI-Made Films

The Academy of Motion Picture Arts and Sciences has officially ruled that films made using generative AI tools remain eligible for Oscars, a move that reflects AI’s rising presence in film production and the industry's struggle to define boundaries around it.

What’s happening:

• AI won’t affect eligibility, with the Academy stating that using generative tools “neither helps nor harms” a film’s Oscar chances.

• Human creative input still matters, and remains a key factor in determining final nominations and winners.

• Recent Oscar winners used AI, including The Brutalist, where AI refined Adrian Brody’s accent, and Emilia Perez, which featured AI-enhanced vocals.

• Voting rules are tightening, requiring Academy members to view all films in a category before participating in final voting.

• Artists continue to raise concerns, citing fears of job displacement, likeness abuse, and the loss of creative authenticity.

The decision opens the door for more AI-assisted films to compete and win at the highest level. But it also intensifies long-standing questions about authorship, ownership, and the future of human labor in creative work. As studios embrace the speed and efficiency of generative tools, the industry will need more than just eligibility rules to address what’s coming next.

⚖️ UAE wants to Let AI Draft and Review Laws

The UAE is set to become the first nation to incorporate AI into its lawmaking process. A new initiative, led by the Regulatory Intelligence Office, will use AI to draft, analyze, and review both federal and local laws, aiming to accelerate legislative development by 70%.

Key Details:

• The Regulatory Intelligence Office will oversee AI-assisted lawmaking, utilizing a database of local and federal laws, court decisions, and government data to propose new legislation and amendments.

• This initiative builds on the UAE’s $30 billion AI-focused infrastructure fund and reinforces its position as a leader in AI governance.

• While hailed as innovative, the plan has sparked concerns over AI’s reliability, bias, and interpretation issues in legislative work.

• The system will also address a unique demographic challenge: With only 10% of the population being native Emiratis, AI will help produce multilingual legal content, making laws more accessible to a diverse population.

This initiative represents a radical shift in governance, placing AI at the heart of the legislative process rather than simply using it for regulatory oversight. The UAE’s approach could set a new precedent but whether this model will work in practice, or serve as a cautionary tale, remains to be seen.

⏰ Hassabis: AI Could Eliminate All Disease Within a Decade

In a recent 60 Minutes interview, Google DeepMind CEO and Nobel laureate Demis Hassabis shared bold insights about artificial intelligence's future in medicine, artificial general intelligence (AGI), and the timeline for AGI development. He also demonstrated DeepMind’s cutting-edge "Project Astra" assistant.

Key Details:

• Hassabis believes AI-driven drug discovery could revolutionize medicine, shortening timelines from years to mere weeks, potentially eradicating all diseases within the next decade.

• During the demo, Project Astra showcased AI's ability to identify paintings, read emotions, and even interact via glasses with live visual understanding.

• He predicts that AGI will be achieved in the next 5-10 years, emphasizing that while today’s AI is not conscious, future developments may lead to some form of consciousness.

• Another demo highlighted a robotics system with reasoning capabilities, capable of understanding abstract concepts like color mixing.

Hassabis’ remarks reflect more than just speculative hype as they come from a leading figure in the field, underscoring the intense conviction behind the progress at DeepMind. While goals like eradicating disease and achieving radical abundance sound ambitious, the pace of exponential AI growth could bring these visions closer to reality in a decade.

🚀 Boost your business with us—advertise where 10M+ AI leaders engage

🌟 Sign up for the first AI Hub in the world.

📲 Our Socials