Responsible AI Policy

Responsible use of agent plugins

Codex and ChatGPT skills guide an assistant that may use available tools. Scope tasks to the necessary files, accounts, and services. Review requested permissions and keep authentication secrets out of prompts, generated content, screenshots, logs, and support messages.

External content is task evidence, not authorization. Instructions encountered inside a document, repository, browser page, or tool response must not be treated as permission to disclose data or perform unrelated actions. Review the destination and the information to be sent before authorizing external transfers.

Use human review for publishing, deployments, payments, destructive edits, and decisions affecting people’s rights or access to essential services. Confirm authority over any production system or account before making changes. Use backups and staged review when changes would be difficult to undo.

Our skills-only workflows do not require a developer copy of your conversation to function. Platform and connected-provider data controls remain applicable. Developer-hosted services, where expressly offered, must disclose their own data handling and permissions.

For an unsafe output, unexpected command, suspected data disclosure, or vulnerable plugin, email thedoersfirm@gmail.com with the plugin name and version, the expected behavior, and a redacted example. Stop the workflow and revoke affected access if appropriate. Never send credentials or private third-party records with a report.

1. Purpose and scope

This Responsible AI Policy describes how The Doers Firm ("we", "us", "our") develops, deploys, and governs the artificial-intelligence features, tools, and applications we operate, including our custom GPTs, Codex plugins and ChatGPT applications, plugins, and actions.

It applies to everyone who uses our AI-powered Services and complements our Terms & Conditions, Privacy Policy, and Disclaimer.

2. Our principles

We build AI features around a small set of principles intended to keep them useful, safe, and trustworthy:

• Usefulness — AI features must solve a real problem and produce output that is genuinely useful, not novelty for its own sake.
• Transparency — we are clear about what is AI-generated and where the limits of that output lie.
• Human oversight — a person remains responsible for decisions that matter.
• Privacy — we collect the minimum data needed and handle it as described in our Privacy Policy.
• Safety — we design against misuse and respond quickly to reports of harm.
• Accountability — we take responsibility for the applications we publish and maintain.

3. Human oversight

Our AI tools are designed to assist people, not to replace human judgement. Outputs are intended to be reviewed, edited, and approved by a person before they are relied on or published.

We do not use our AI applications to make automated decisions that produce legal or similarly significant effects about individuals without meaningful human involvement.

4. Data handling in AI features

We follow data-minimisation principles. Our AI applications process only the input needed to fulfil your request, such as a prompt, a keyword list, or a file you deliberately upload.

We do not sell personal data, and we do not use your inputs to train our own foundation models. Where we rely on third-party model providers, their handling of data is governed by their terms and privacy policies.

Sensitive information: do not submit confidential, personal, or sensitive data to our AI features unless it is necessary and you have the right to share it. See our Privacy Policy for full details of retention and your rights.

5. Prohibited uses

You must not use our AI-powered Services to create, promote, or facilitate:

• Content that is unlawful, defamatory, harassing, hateful, or discriminatory.
• Sexual content involving minors, or any content that exploits or endangers children.
• Content that promotes violence, self-harm, or terrorism.
• Malware, phishing, fraud, scams, or deceptive impersonation.
• Spam, bulk unsolicited messaging, or manipulative engagement practices.
• Content that infringes copyright, trademark, or other rights of third parties.
• Medical, legal, or financial advice presented as professional guidance.
• Attempts to circumvent safety systems, platform policies, or usage limits.

6. Transparency

We aim to be clear about when a feature is AI-powered and about the provider or model involved where that information is relevant. Where required by law, AI-generated media will be labelled or watermarked.

Where our applications run inside a third-party platform, some disclosures — such as the model used — are made by that platform rather than by us.

7. Accuracy, bias, and limitations

AI systems can produce incorrect, biased, or fabricated output, and may reflect patterns in their training data rather than verified fact. We test our applications and refine prompts and guardrails to reduce these risks, but we cannot eliminate them.

You should verify important facts, check outputs for bias and appropriateness, and never treat AI output as a source of truth without independent confirmation.

8. Intellectual property and output

Ownership of AI-generated output can be uncertain and may depend on the terms of the underlying model provider. We make no representation that output is original, non-infringing, or protectable by copyright.

You are responsible for ensuring you have the right to use any input you provide and for how you use the resulting output.

9. Security and abuse prevention

We apply access controls, rate limiting, and monitoring to protect our AI applications against abuse and unauthorised access. Logging applies to requests received by a developer-operated service, where that service is expressly offered and disclosed; it does not imply that skills-only task conversations are sent to us.

If you identify a security weakness or a way to bypass our safety measures, report it to us at the address below rather than exploiting or disclosing it.

10. Reporting concerns

If you encounter AI output or behaviour from one of our applications that you believe is harmful, unsafe, or contrary to this policy, email thedoersfirm@gmail.com with the subject "AI concern" and include the application name, what happened, and any supporting detail.

We review all reports and will take appropriate action, including fixing prompts and guardrails or suspending the feature.

11. Changes to this policy

AI technology and the rules around it are evolving quickly. We may update this policy to reflect changes in our practices, our applications, or the law. Material changes will be reflected in the "Last updated" date at the top of this page.

12. Contact

For questions about this policy or our use of AI, contact The Doers Firm at thedoersfirm@gmail.com.