We are following Russel and Norvig’s definition, as mentioned in the introduction.

Joint work with my wonderful colleagues:

@jennwv.bsky.social
Dan Weld
Saleema Amershi
@erichorvitz.bsky.social
@adamfourney.bsky.social
Hussein Mozannar
Victor Dibia

#AIAgents #LLMs #TechNews

5/

We're calling on researchers and practitioners to prioritize these issues and enhance transparency, control, and trust in AI agents! 📄 Read full details at microsoft.com/en-us/resear...

4/

Why does this matter?

Without proper grounding, we risk safety failures, loss of user control, and ineffective collaboration. Trust and transparency in AI systems hinge on addressing these challenges. We supplement each challenge with examples.

3/

Some challenges focus on how agents can convey necessary information to help users form accurate mental models (A1-5). Others address enabling users to express their goals, preferences, and constraints to guide agent behavior (U1-3). We also focus on many overarching issues (X1-4).

2/