How does the world feel about the possibility of understanding other species?

Findings from our Global Dialogue on AI and interspecies communication.

Nov 5 • 

CIP

1

October 2025

When AI Acts For You (Or As You)

Early findings from our sixth Global Dialogues on frontier AI agents.

Oct 17 • 

CIP

3

2

1

There will soon be millions of AI agents working on our behalf for work, commerce, and democracy.

How do we ensure people’s preferences and values are faithfully represented?

Oct 10 • 

CIP

8

3

Why You Should Care About AI Evaluations

The real world is complex, and we need AI evals that capture that complexity.

Oct 6 • 

CIP

 and 

Faisal Lalani

2

September 2025

Notes on building collective intelligence into evals

Incorporating collective intelligence into context-specific evaluation

Sep 29 • 

Divya Siddarth

 and 

CIP

10

3

2

A global snapshot of trust and AI

Findings from our Global Dialogues data, which encompasses ~4,000 participants across four rounds, reveal that trust in traditional institutions is low…

Sep 23 • 

CIP

4

1

People are relying on AI for emotional support. This has global implications for trust and governance.

Some findings from our Global Dialogues on Human-AI relationships

Sep 16 • 

CIP

2

People are starting to believe that AI is conscious.

And this is already affecting the world, regardless of whether or not AI ever actually is.

Sep 8 • 

CIP

12

10

6

Frontier AI Agents with collective input

What does the world want from future AI agents? Announcing our partnership with the Industry-Wide Deliberative Forum, and our latest Global Dialogues…

Sep 4 • 

CIP

4

1

1

August 2025

Translating global public input into AI evals

Early insights from Global Dialogues

Aug 25 • 

CIP

5

2

Collective Intelligence with LLMs

From chain-of-thought to chain-of-deliberation

Aug 21 • 

CIP

13

4

4

LLM Judges Are Unreliable

How Positional Preferences, Order Effects, and Prompt Sensitivity Undermine Reliability in AI Judgments

Aug 15 • 

CIP

6

2

1