1. Which Tools Use AI
Tools labelled "AI" in our catalogue use one or more third-party large-language-model (LLM) APIs. Examples include the AI content generators, AI summarisers, AI image generators, and AI-assisted SEO recommendation tools. Where a tool uses AI, the tool page indicates so explicitly. Other tools (calculators, converters, hash generators, EXIF readers, etc.) do not use AI.
2. Which Providers We Use
We route AI requests to one or more of the following providers, depending on the tool and your settings:
- Anthropic (Claude) — primary text-generation provider.
- OpenAI (GPT) — secondary text-generation and image-generation provider.
- OpenRouter — routing layer to additional LLM providers, when configured.
Each provider's usage policies bind you in addition to ours. The complete current sub-processor list, with regions and safeguards, is on our Sub-processors page.
3. What Happens to Your Inputs
- The prompt and any inputs you submit to an AI tool are sent over TLS to the chosen provider's API.
- We use the providers' API offerings (not their consumer products). Anthropic and OpenAI contractually commit to not training on data submitted via the API by default.
- Where the provider supports it, we use zero-retention mode, so prompts are not stored on the provider's side beyond the duration of the request.
- We do not send your account profile, IP address, or browsing history to AI providers along with the prompt.
- We may log a hashed identifier of your request for rate-limiting, abuse detection, and billing reconciliation. We do not log the content of prompts in our application logs.
4. Do Not Submit Sensitive Data
You must not paste the following into any AI tool:
- Protected health information (PHI) within the meaning of HIPAA.
- Payment card data (PCI), full card numbers, CVVs, or full bank-account numbers.
- Government identifiers (Social Security numbers, driver's-licence numbers, passport numbers).
- Trade secrets, classified information, or attorney-client privileged material.
- Personal data of others for which you do not have a lawful basis to process.
If you submit such data anyway, you do so at your own risk. We do not represent that the providers' systems meet HIPAA, PCI-DSS, ITAR, or similar specialised regulatory standards.
5. Output Quality and Limitations
AI outputs are produced by statistical prediction over training data and can be:
- Inaccurate or "hallucinated" — presenting plausible-sounding but false statements as fact.
- Out of date — reflecting training data with a cut-off prior to today.
- Biased — reflecting biases in training data.
- Inconsistent — the same prompt may produce different outputs on different runs.
- Inadvertently infringing — reproducing copyrighted, trademarked, or proprietary text or imagery from training data.
You are responsible for reviewing, fact-checking, and editing AI outputs before relying on them, publishing them, or making decisions based on them. Treat AI outputs as a draft, not a finished work product.
6. Ownership of Outputs
As between you and us, you own the outputs an AI tool generates from your prompts, subject to:
- The terms of the underlying third-party provider (e.g. Anthropic Commercial Terms, OpenAI Business Terms).
- Any third-party rights in pre-existing material reflected in the output (an AI output is not automatically free of copyright if it reproduces copyrighted material from training data).
- Restrictions in our Terms of Service and our Acceptable Use Policy.
- The fact that, under current U.S. Copyright Office guidance, purely AI-generated works without sufficient human creative contribution are not eligible for copyright protection in the United States.
7. Disclosure Obligations
You should disclose your use of AI in any context where non-disclosure would mislead a reasonable user, including:
- Academic submission (most institutions require disclosure or prohibit AI use entirely).
- Journalism (most outlets require disclosure of AI assistance).
- Regulated professional advice (legal, medical, financial — many jurisdictions require disclosure or prohibit unsupervised AI use).
- Reviews, testimonials, or endorsements (FTC requires disclosure of material connections, including AI generation).
Where applicable law requires AI-generated content to carry a label or watermark (e.g. EU AI Act Art. 50, certain U.S. state laws on political ads, deepfake disclosure laws), you are responsible for adding the required disclosure to outputs you publish.
8. High-Risk Uses Are Prohibited
You may not use AI outputs from the Service:
- To make automated decisions that produce legal or similarly significant effects on individuals (Art. 22 GDPR), without independent human review.
- For any "high-risk AI system" application as defined in Annex III of the EU AI Act (Reg. (EU) 2024/1689), including biometric identification, critical infrastructure operation, education and vocational training admissions, employment screening, eligibility for essential services, law enforcement, migration and border control, and administration of justice.
- To generate election disinformation, deepfake political content, or other content that violates laws regulating AI-generated material (e.g. California AB 730/2655, Texas SB 751).
- To generate non-consensual sexual imagery of identifiable real persons, regardless of jurisdiction.
- To diagnose, treat, or make decisions about specific medical, legal, or financial matters affecting third parties without involving an appropriately licensed professional.
9. Children and AI
Our AI tools are not directed to or appropriate for children under 16. Outputs may include themes, language, or recommendations that are inappropriate for minors. If you are under 16 (or under 13 in the United States), do not use AI tools without parental consent and supervision.
10. AI Training Data Position
We do not use your prompts, outputs, or any other user data to train AI models. We do not sell or share user inputs with third parties for the purpose of training their models. Where a future product release contemplates such use, we will obtain your explicit opt-in consent first and provide a clear opt-out mechanism.
11. Provider-Side Logging
Even when we configure zero-retention mode, AI providers may temporarily process and (very briefly) cache prompts in memory or cache for safety filtering, abuse detection, and infrastructure operation. Each provider's privacy practices govern that processing. By using an AI tool, you authorise that limited provider-side processing as part of fulfilling your request.
12. Changes to AI Providers and Models
We may add, remove, or substitute AI providers and models at any time, including in response to provider availability, performance, cost, or compliance with our standards. The current list is maintained on the Sub-processors page. Material changes affecting your privacy are also reflected in our Privacy Policy.
13. No Warranty
AI outputs are provided "AS IS" with no warranty of any kind, express or implied. The Disclaimer in our Terms of Service and our General Disclaimer apply in full to AI features.
14. Reporting AI Misuse
To report misuse of an AI tool (e.g. prompt-injection attempts, evasion of safety filters, generation of illegal content), email info@baaed.com with the subject line "AI Safety Report".
15. Contact
Questions about this AI Policy: info@baaed.com.