Legal

Legal, Disclaimer & Privacy

What BlacklistAI is, what it is not, where our information comes from, and how we treat your data.

Listy, a dark blacklist of flagged threats holding a pen

Last updated: 20 November 2026. Operated as blacklistai.org (free beta).

1. Main disclaimer

BlacklistAI is an information and research service about publicly documented AI failures, misuse and security weaknesses. Everything on this site — the Overview, Threat Feed, Threat Map, Live Radar, Reports, Trends, Assessment, Safety Check, Stress Test, Secret Tripwire, Trust Badge and API — is provided "as is" and "as available", for general information only, without any warranty of accuracy, completeness, timeliness or fitness for a particular purpose.

Using this site does not make your AI system safe, compliant or certified. You remain fully responsible for decisions about your own systems.

IN PLAIN WORDS:We show what has gone wrong for others and how people usually defend against it. We don't guarantee it will protect you.

2. Not legal, security or compliance advice

Nothing here is legal, financial, professional security, insurance or regulatory compliance advice, and no client, advisory or fiduciary relationship is created. Defense guidance (including references to OWASP Top 10 for LLM Applications, MITRE ATLAS, NIST AI RMF and GitHub security advisories) is general, category-level guidance — not a tested patch for your specific system — unless a record states an exact advisory match.

Reports, assessments, scores and badges must not be presented as an audit, certification or proof of compliance with any law or standard (for example the EU AI Act, GDPR or sector rules). Consult qualified professionals before acting.

3. Data sources & attribution

Documented incidents come from third-party public archives, chiefly the AI Incident Database (AIID) and the AIAAIC Repository, plus a small set of curated public cases. Each incident links to its original source; those sources and their owners hold their own rights and terms.

Frameworks and advisories are referenced from OWASP, MITRE ATLAS, NIST and the GitHub Advisory Database. We are not affiliated with or endorsed by any of these organizations.

Map locations marked "approximate" use a country's centre point, not the real site. Severity scores, sector, provider and model labels are our own classifications and may be wrong or inferred. When our database is offline, pages show a saved snapshot labelled with its last-updated date.

4. Named companies & AI products

Company, product and model names (for example OpenAI/GPT, Anthropic/Claude, Google/Gemini) appear only because they are named in public source records. Rankings and counts reflect how often a name appears in documented records — not confirmed exploitation, liability, affected versions, or that one product is less secure than another. Trademarks belong to their owners; no endorsement or affiliation is implied.

5. Live Radar (unconfirmed signals)

Live Radar collects early-warning headlines and posts from public sources such as Google News, Reddit and Hacker News. These are unverified, automatically tagged, and kept separate from documented incidents. They may be inaccurate, rumour, or later disproved. Do not rely on them or repeat them as fact.

6. AI-generated analysis

The Assessment and deep checks use AI models to summarise and rank information. We restrict citations to real records in our archive, but AI output can still be incomplete or wrong. Treat it as a starting point for your own review.

7. Free tools

The Safety Check, Stress Test and Secret Tripwire (canary) tools are educational. A "pass" does not mean your prompt or system is secure. Only test systems you own or are authorised to test. Text you paste may be processed on our servers to produce a result; do not paste passwords, keys or personal data.

8. Accounts, credits & payments

You can sign in with email or Google. Free welcome credits power certain deep checks; they have no cash value, cannot be transferred, and may be changed or withdrawn during the beta. Detailed defense blueprints are coming soon.

9. Acceptable use

Don't use this site to attack, exploit or harass anyone, to break laws, to scrape or resell the service in bulk, to overload our systems, or to bypass access limits. Submissions and reports you send must be truthful and must not contain other people's private data. We may remove content or suspend accounts that break these rules.

10. Limitation of liability

To the fullest extent the law allows, BlacklistAI and its operators are not liable for any direct, indirect, incidental or consequential loss — including security breaches, regulatory penalties, lost profits or data loss — arising from your use of, or reliance on, this site, its data, analysis, tools or third-party links. Some jurisdictions don't allow certain limitations, so parts of this may not apply to you.

11. Corrections & removal requests

If you believe a record, classification or company mention is inaccurate, email support@blacklistai.org with the incident link and the original source. We review requests and correct, relabel or remove content where the public evidence supports it.

12. Changes & contact

We may update this page as the service changes; the date at the top shows the latest version. Continued use means you accept the current version.

Contact: support@blacklistai.org · Back to Overview

About trust badges

A "Self-assessed with BlacklistAI" badge means the owner of a website or AI system used BlacklistAI's free tools to check their own system. BlacklistAI has not audited, tested, certified or approved that system, and the badge is not a guarantee of safety.

Data sources & licensing

AIAAIC Repository — Some incident data comes from the AIAAIC Repository (https://www.aiaaic.org). AIAAIC content is licensed under CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0/). BlacklistAI has adapted it (shortened, combined fields, categorised and located incidents). Our adapted version of AIAAIC data is shared under the same CC BY-SA 4.0 licence.

AI Incident Database — Some incident data comes from the AI Incident Database (https://incidentdatabase.ai), run by the Responsible AI Collaborative, and is used in line with its Terms of Use (https://incidentdatabase.ai/terms-of-use/). Each incident links to its original record.

Live Radar — Live Radar items are unverified early signals collected automatically from public feeds (Google News, Reddit, Hacker News, GitHub Security Advisories, arXiv). They are not confirmed incidents. Each item links to its original source.

Community reports — Threats on the Threat Feed are submitted by users and are not verified by BlacklistAI.

Methodology — BlacklistAI groups reported incidents by system type and failure mode. Counts show reported incidents only, not how often problems really happen. BlacklistAI does not rank AI companies, providers or models, and nothing on this site is a safety rating of any company or product.

Corrections — Spotted a mistake? Use the "Report an error" link on any incident, or email tannouswalid@gmail.com.

Privacy policy

Last updated: 20 November 2026. Contact: tannouswalid@gmail.com.

Account: If you sign up, we store your email address and display name, and a password managed securely by our authentication provider (Supabase), or your Google name and profile picture if you sign in with Google. We also store your credit balance.

Threat Feed: Reports and votes you submit are stored with your user ID.

Threat Assessment: The industry, AI provider and description you enter are sent securely to our server and to an AI model through the Lovable AI Gateway to create your assessment. We do not save them in our database, and we request that the AI provider does not store them.

Safety Check, Stress Test, Secret Tripwire and Trust Badges: Basic checks run entirely in your browser. Deep checks (which use credits) send your text to our server and an AI model to produce a result; we do not save it in our database.

Browser storage: We save a guest credit counter and, if you are logged in, your login session in your browser's local storage. We use one cookie ("sidebar:state") only to remember whether the menu is open.

Analytics: We use Lovable's built-in visitor statistics, which count visits by country, browser and referring site without tracking cookies. We do not use Google Analytics, Meta Pixel or any other tracking scripts.

Service providers: Lovable (hosting and AI gateway), Supabase (database and login), Google (only if you choose Google sign-in).

We do not sell your personal data. To see or delete your data, or delete your account, email tannouswalid@gmail.com.

Terms of use

BlacklistAI is provided for information and education. Incident information comes from third-party sources and may contain errors. Nothing on this site is legal, security or professional advice. Do not use the tools to attack or harm any system. Use of the site means you accept these terms.