Consulting & Applied Research
Responsible use
Use data, AI, automation and research with permission, judgement and appropriate oversight.
Start with authorised access
Use data, software, accounts and APIs only where you have the necessary permissions. Respect the applicable platform terms, access limits, intellectual property rights and confidentiality obligations. Access to a tool or dataset does not by itself grant permission for every possible use.
Share only the information needed for the task. Avoid placing passwords, private keys, sensitive personal information or confidential client material into tools that have not been approved to receive it. Confirm appropriate handling arrangements before connecting a service to business systems.
Keep people accountable
AI outputs can be inaccurate, incomplete, outdated or inappropriate for the context. Check important claims, calculations, citations and generated code before relying on them. A polished result is not evidence that its underlying facts are correct.
Apply human review and suitable testing before publishing content, making external changes or using an output in a consequential decision. High impact uses involving health, safety, finances, employment or legal rights require oversight appropriate to the domain. Automation should have clearly assigned responsibility and a way to identify and correct errors.
Respect people and systems
Do not use our website, materials or services to gain unauthorised access, bypass safeguards, impersonate people, infringe rights or send abusive or unsolicited bulk messages. Automated publishing does not remove responsibility for accuracy, permissions or the rules of each destination.
Review integrations for security and limit their access to what the task requires. Do not assume that a portfolio demonstration establishes suitability for your production environment.
Interpret research carefully
Prediction market material is intended for adults aged 18 or over. A venue may impose a higher age threshold or additional location and eligibility rules. Check those rules yourself and use a venue only where you are legally eligible.
Research can be wrong, prices can change and participation can result in financial loss. Historical examples are not current recommendations. Verify the exact contract, outcome, observation time and underlying source independently. Do not treat a model estimate or Model Gap as a guaranteed result, and do not commit money you cannot afford to lose.
Raise a concern
Report a suspected error, misuse or security concern to support@precisionalgorithms.com. Describe the issue and the affected page or service without including passwords, private keys or unnecessary personal information.