Wharton labs open-sources AIBO for statistically valid AI behavior experiments
emollick · x · 2026-07-29
Wharton Generative AI Labs released AIBO, an open-source AI Behavioral Observatory for running statistically valid, controlled experiments on AI systems at scale.
- Researchers can define control and treatment prompts, run many trials concurrently, score outputs, and analyze results statistically.
- The team says the tool shortened the distance between a research question and evidence, and they have already used it in their own studies.
- It is intended for behavioral research on how AI responses change under different prompts, not just one-off prompting.
More from Companies & People
- Eragon adds routing across 350+ models via Merge API for agent orchestration — shensi · 2026-07-30
- alphaXiv adds Hugging Face login in a new research-discovery partnership — huggingface · 2026-07-30
- 8090 says it raised a Series A as Chamath frames AI as a threat to outsourced hours — vaibhavbetter · 2026-07-30
- Decoder Ring Needed: Navigating Google's Confusing AI Product Naming — tristanbob · 2026-07-30
- AI employee startup hires Robbie O'Connor after a year watching the market — alex_verem · 2026-07-30
- An MCP feedback tool could file user feature requests — but may mostly collect compliments — hermjohnson · 2026-07-30