Anthropic
Claude
Long-context reasoning with a published safety policy.
Impact Indicators
Job Displacement
No verified Claude-specific job-loss data exists. Anthropic’s own usage research measures occupational exposure and automation patterns, while independent labor research has not established broad Claude-attributed employment losses.
Privacy & Surveillance
Anthropic discloses that consumer chats may be used for model improvement when users allow it, with some opted-in data retained for up to five years; commercial API inputs and outputs generally follow a shorter default retention period.
Water Usage
No measured Claude-specific water-use figure exists. AWS infrastructure serving Anthropic operates within regional water systems and facility-level limits, but those figures cannot be attributed directly to Claude.
Energy & Carbon
No independently verified Claude-specific energy or emissions figure exists. Anthropic uses large-scale compute across AWS and other cloud infrastructure, while broader data-center research shows rapidly rising electricity demand from AI workloads.
U.S. reporting requirements do not yet provide standardized, model-level disclosure for many AI human and environmental impacts. When U.S. evidence is incomplete, DoctorKnow uses verifiable EU regulatory disclosures, public-sector data, independent research, and reputable journalism to provide the strongest evidence currently available.
Evidence scope is identified as model-specific, provider-level, facility-level, regional/grid, or industry-level.
Scores use a 0–100 scale. Higher scores indicate stronger performance or lower concern; 100 is the best possible rating.
Reasoning
96
Coding
93
Writing
97
Analysis
95
Speed
68
Scores use a 0–100 scale, where 100 represents the strongest outcome and 0 the weakest.
What Claude does
Claude is built by Anthropic and is positioned for writing, analysis, research and legal. Our readings score it strongest on long-form writing and multi-step reasoning.
Its weakest measured axis is response speed (68/100), so plan around that when the work depends on it. Access is paid, and the weights are closed, so you can only reach it through the vendor's own service.
Ethical stance
Transparency reads 84/100. Documentation, evaluations and model details are unusually open for the category.
Labor and job-displacement posture reads 71/100. Data-worker conditions and displacement effects are addressed rather than ignored.
Regulatory posture reads 88/100. Claude's vendor engages constructively with binding AI rules.
Privacy and data handling
Privacy and data handling reads 76/100. Retention is short or optional, training on your prompts is off by default, and enterprise terms are clear.
Because the weights are closed, every prompt is processed on vendor infrastructure — treat regulated or confidential data accordingly.
Energy, carbon and water
Carbon footprint reads 66/100. The energy mix is mixed, and disclosure covers some but not all serving regions.
Water and local impact reads 62/100. Some sites use evaporative cooling in regions with moderate water stress.
Speed and efficiency also matter here: a 68/100 speed reading means longer runs per answer, which raises the energy cost of high-volume use.
Diagnostic notes
Publishes responsible scaling policy and supports binding safety regulation.
Readings are DoctorKnow.ai's own 0–100 assessments, compiled from vendor documentation and public reporting, and revised as new disclosures land. They are a comparison aid, not a certification.