🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Last month we launched our Anthropic Economic Index, to help track the effect of AI on labor markets and the economy. Today, we’re releasing the second research report from the Index, and sharing several more datasets based on anonymized Claude usage data.

We’re sharing a new report and several new datasets as part of our Economic Index! Highlights include: - Changes in usage since Claude 3.7 Sonnet - Automation/augmentation broken down by task - Extended thinking mode usage by task - A new bottom-up taxonomy of Claude usage 🧵1/

Excited to share that our paper "Bridging the human–AI knowledge gap through concept discovery and transfer in AlphaZero" is now out in PNAS! With Nenad Tomasev, Tom McGrath, Demis Hassabis, Ulrich Paquet, Been Kim 🎉 📄 doi.org/10.1073/pnas.2…

New Anthropic research: How university students use Claude. We ran a privacy-preserving analysis of a million education-related conversations with Claude to produce our first Education Report.

We just released our Anthropic Education Report analyzing ~1M anonymized student conversations on Claude using our privacy-preserving Clio tool. Kunal Handa and I discovered that students use AI ~equally across 4 learning modes, STEM dominates usage (36.8% CS), and students

AI will fundamentally transform how we learn. Our new report on education use cases of Claude shows the early signals of this shift and what it means for students, educators, and institutions.

Anthropic continues to put out solid studies of how Claude is used, this time in universities It shows both AI perils & promise. Students are using it for all sorts of purposes, with getting direct answers making up a lot use, but many cases of AI supporting their learning; too.

Do LLMs build structured internal models of the world? How do we formalize, extract, and assess these models? If you’re working on these types of questions, consider submitting to our workshop on Assessing World Models @ ICML 2025! Deadline is May 20

New Anthropic research: AI values in the wild. We want AI models to have well-aligned values. But how do we know what values they’re expressing in real-life conversations? We studied hundreds of thousands of anonymized conversations to find out.

Really proud and excited to release this work on empirically measuring AI values “in the wild” — understanding, analyzing and taxonomizing what values guide model outputs in real interactions with real users. There is a lot of work on training models to follow particular

Excited to share our new research with Saffron Huang on studying Claude's values in real-world conversations! We've created a comprehensive taxonomy of AI-expressed values based on interactions in the wild.

I'll be presenting our work "Eliciting Human Preference with Language Models" at ICLR! Come catch my poster Thursday 4/24 at 10AM → iclr.cc/virtual/2025/p… Also DM me if you're interested in world models, interpretability, personalized interaction, or just general chatting!

Want to study the economic impact of AI and influence the policy choices a frontier lab makes? I'm building a team to advance Anthropic 's Economic Index & other ~special projects. Lots of fun! Economist: job-boards.greenhouse.io/anthropic/jobs… Data Scientist: job-boards.greenhouse.io/anthropic/jobs…

🧵✨🙏 With the new Claude Opus 4, we conducted what I think is by far the most thorough pre-launch alignment assessment to date, aimed at understanding its values, goals, and propensities. Preparing it was a wild ride. Here’s some of what we learned. 🙏✨🧵

Happy to announce the first workshop on Pragmatic Reasoning in Language Models — PragLM @ COLM 2025! 🧠🎉 How do LLMs engage in pragmatic reasoning, and what core pragmatic capacities remain beyond their reach? 🌐 sites.google.com/berkeley.edu/p… 📅 Submit by June 23rd

New Anthropic Research: How people use Claude for emotional support. From millions of anonymized conversations, we studied how adults use AI for emotional and personal needs—from navigating loneliness and relationships to asking existential questions.

Important and timely work led by Miles McCain and 309 others et al. on affective conversations with Claude—how people turn to AI for support, advice, and companionship, and how Claude responds in these moments.

New Paper: Can we collect human chains-of-thoughts by asking them to think out loud? In our new paper we automate and study this protocol with 5,000 human reasoning traces from 640 people solving Countdown problems. 1/5

Announcing the Anthropic Economic Futures Program—our latest commitment to understanding AI's impacts on work and the economy. The program will support new research and actionable policy solutions to address the workforce impact of AI.

Thrilled to join the UMich faculty in 2026! I'll also be recruiting PhD students this upcoming cycle. If you're interested in AI and formal reasoning, consider applying!