Question 1

What is the AI Privacy Scorecard?

Accepted Answer

A free interactive tool that compares 15 AI assistants — ChatGPT, Claude, Gemini, Microsoft Copilot, Jarvis, Mistral Le Chat, Apple Intelligence, Perplexity, Notion AI, Cluely, Highlight AI, Raycast AI, Aleph Alpha, DeepL Write, and Wispr Flow — across 9 privacy dimensions and produces a personalized recommendation based on your ranked priorities.

Question 2

What 9 privacy dimensions does the scorecard cover?

Accepted Answer

EU data residency, training opt-out, encryption at rest (AES-256-GCM and similar), GDPR posture (Data Processing Agreement plus third-party audit), EU AI Act readiness (Article 50/53/55), audit logs (SIEM-exportable), SSO support (SAML and SCIM), retention controls with explicit deletion SLA, and bring-your-own-key (BYOK) support.

Question 3

Which AI tool has the best privacy?

Accepted Answer

Mistral Le Chat (Paris, EU-default) and Aleph Alpha (Germany, BSI C5 certified) score highest in raw points. Microsoft Copilot leads for Microsoft 365 customers thanks to the EU Data Boundary commitment. Apple Intelligence wins on on-device residency. Jarvis ranks well on EU residency plus AES-256-GCM plus no-training defaults, and poorly on BYOK, which it does not support. The right answer depends on your weighted priorities — that is exactly what the interactive tool computes.

Question 4

Is the AI Privacy Scorecard biased toward Jarvis?

Accepted Answer

The tool is published by Jarvis (one of the 15 rated tools) and every score is reproducible from each vendor's linked public policy. Jarvis is honest about its weak spots: it ranks well on EU-default residency, AES-256-GCM token encryption, and no-training defaults, but scores low on BYOK because it has none, and lower on SAML/SCIM SSO and SIEM-style audit logs. See the methodology section for the per-cell sourcing.

Question 5

How does the recommendation engine work?

Accepted Answer

You pick 3-5 tools and rank your top 5 privacy priorities. Your top priority counts 5×, your second counts 4×, then 3×, 2×, and 1×. Dimensions you did not pick still contribute at weight 1 — they matter, just less. The tool with the highest weighted total wins and the per-dimension breakdown is shown side-by-side so you can see exactly why.

Question 6

Can I share my AI Privacy Scorecard result?

Accepted Answer

Yes. Every result has a unique URL — copy it, tweet it, email it, or use the embed code to drop the scorecard into an article or blog post. The shared link always reopens the exact same comparison. Result pages are noindex so they do not compete with the main tool page for search rankings.

AI Privacy Scorecard: 15 Assistants Compared.

What should you check first?