Scorecard_grade·saferai-2025-10|sid_A4XoubikkQ|risk-identification·Record·Profile

Scorecard: saferai 2025-10-01 scored Google DeepMind on Risk Identification = Very Weak

Verdictconfirmed95%

1 check · 9/21/2026

1 → confirmed

Our claim

entire record

Snapshot: saferai-2025-10
Entity: Google DeepMind
Dimension Slug: risk-identification
Dimension Label: Risk Identification
Dimension Weight: 0.25
Score Numeric: 22
Score Letter: Very Weak
Score Raw: 22%

Source evidence

1 src · 1 check

ratings.safer-ai.org/comparison/resource

confirmed95%Haiku 4.5 · 9/21/2026

NoteThe source is a comparison table from SafeRAI's 2025 scorecard (dated October 2025, matching the claimed publishedAt date of 2025-10-01). The table explicitly shows Google DeepMind in the Risk Identification row (section 1) with a score of 22%. This matches the claimed scoreRaw of 22% and scoreNumeric of 22. The 'Very Weak' letter grade is consistent with a 22% score on a typical grading scale. All key fields are confirmed: publisher (saferai), publishedAt (2025-10-01), entity (Google DeepMind), dimension (Risk Identification), and the numeric score (22%).

Case № saferai-2025-10|sid_A4XoubikkQ|risk-identificationFiled 9/21/2026Confidence 95%