Compassion Benchmark

Score history

OpenAI

27.5developingRank #42
15 scored eventsMay 20, 2026June 27, 2026
View entity page →
27.5

Score timeline

Scored
◆ Hold27.5developing

42-state attorney general subpoena and Florida criminal investigation remain pre-adjudication — neither has concluded; score holds at 27.5.

Scored
◆ Hold27.5developing

42-state attorney general subpoena and Florida criminal investigation both pre-adjudication — neither has concluded; score holds at 27.5.

Scored
◆ Hold27.5developing

42-state attorney general subpoena issued June 13; Florida criminal investigation active; both are investigations, not convictions — score holds under the pre-…

Scored
◆ Hold27.5developing

42-state AG subpoena + Florida criminal investigation + Florida civil suit carry forward; pre-adjudication discipline holds; EU AI Act Aug 2 now 40 days away.

Scored
-0.327.2developing

42-state AG subpoena issued five days after IPO filing; Florida civil suit and criminal investigation reference active; pre-adjudication discipline holds.

Scored
-4.722.8developing

42-state AG coalition opened formal investigation; NY AG subpoena served June 12 covering sycophancy, engagement design, health data, and vulnerable users.

Scored
◆ Hold27.5developing

42-state AG subpoena pre-adjudication; ACC/AWR sub-threshold accumulation documented; contrasts with adjudicated K.G.M. verdict.

Scored
◆ Hold27.5developing

42-state AG subpoena is pre-adjudication; sycophancy-as-design-flaw framing documented; ACC/AWR sub-threshold accumulation active.

Scored
◆ Hold27.5developing

42-state AG subpoena names sycophancy a design flaw; pre-adjudication allegation, sub-threshold at −1.6.

Scored
◆ Hold27.5developing

Great American AI Act at procedural stage; Florida child-safety suit filed but unadjudicated.

Scored
◆ Hold27.5developing

DoD classified-deployment guardrails rely on reference-to-law with government interpretive discretion; offset by Frontier Governance Framework alignment with E…

Scored
◆ Hold27.5developing

DoW agreement embeds red lines (no mass surveillance, no autonomous weapons); net-neutral under symmetric MILITARY-AI-BY-CONTRACT-GOVERNANCE hold.

Scored
◆ Hold27.5developing

Frontier Governance Framework published; structural mission-safety word deletion and front-group funding confirmed in same window; INT enters Critical.

Scored
-1.925.6developing

Tumbler Ridge failure-to-report leadership override; front-group funding confirmed; cumulative -3.6 since May 5.

Scored
-1.725.8developing

OpenAI accepted Pentagon contract terms that Anthropic refused — the same terms Altman verbally endorsed refusing — creating a documented gap between stated sa…