Our generative AI policy

This Generative AI Policy (Policy) sets out when AI use must be declared to Lloyd’s Register Foundation (us or we or our) in relation to grant applications and funded work, what is expected from applicants and grant holders, and how to ensure outputs are accurate, transparent and secure.

Responsible use of generative AI in grant applications and funded work 

We recognise that generative artificial intelligence, often called generative AI, can support research, innovation and the delivery of grant-funded work. 

We allow the responsible use of generative AI in grant applications and funded activities. We do not want to stop applicants or grant holders from using useful tools. However, we expect generative AI to be used in an accurate, transparent, ethical and lawful way. 

This Policy explains what we mean by generative AI, when you need to tell us about its use and what we expect from applicants and grant holders. 

What we mean by generative AI 

Generative AI means systems that can create new content. This may include text, images, audio, video, code or other material. 

Generative AI tools work by using patterns learned from large amounts of data. Examples include ChatGPT, Microsoft Copilot and Google NotebookLM. 

This Policy does not apply to standard tools that use fixed rules or simple models and do not create new original content. Examples include spell checkers, translation tools, transcription tools, speech recognition tools and similar systems. 

When you need to tell us about AI use 

You must tell us if you use generative AI to create or support any original content for: 

• A grant application 

• A grant-funded activity 

• A report to the Foundation 

• Images, video, code, text or other materials produced as part of Foundation-funded work 

You do not need to tell us about standard tools as set out above. 

How we will ask you to confirm this 

Grant applicants will be asked to confirm that they have read, understood and agree to follow this Policy as part of the application process. 

This Policy is also applicable to current grant holders . We may include questions about AI use in grant monitoring and reporting. 

Failure to comply with this Policy may be treated as a breach of the grant agreement. The Foundation reserves the right to suspend or terminate funding, require repayment of grant funds, or take other action it considers appropriate where a grant holder has failed to use generative AI responsibly or has not declared its use as required. 

Our expectations 

You are responsible for how you use generative AI. We accept no liability for any loss, claim or damage arising from the use of generative AI in connection with a grant application or funded activity. 

You may use generative AI to support your work, but your application, report or funded activity must remain accurate, honest and based on work that you and your partners can deliver. 

You must use generative AI in a responsible way including the following: 

• Review and verify all content created with generative AI before you submit or use it 

• Make sure the content is accurate, relevant and complete 

• Remove or correct any content that is false, misleading, unclear or unsupported 

• Understand the content you submit to us 

• Make sure your application or report reflects your own plans, evidence, expertise and capacity 

• Avoid using AI to make your organisation, team or project appear more developed or capable than it is 

• Protect confidential information, personal data and unpublished intellectual property 

• Follow your own organisation’s data protection, confidentiality and information security policies 

• Follow all relevant laws and regulations 

• Ensure that the use of generative AI does not infringe any third-party intellectual property rights or create any liability for us 

• Consider whether the AI tool may store, reuse or share the information you enter into it 

• Consider whether AI outputs may contain bias and take reasonable steps to identify and reduce this risk 

AI tools can make mistakes. They may produce information that sounds correct but is false, incomplete or out of date. They may also reflect bias from their training data, model design, language coverage, cultural context, feedback loops or human labelling choices. You must check AI outputs carefully and should not rely on them without human review. 

Where AI-generated content forms part of a funded output, the grant holder must ensure that our rights to use, publish, adapt and share that output are not restricted or compromised by the terms of any AI tool or platform used. 

Grant holders must not use generative AI tools whose terms of service would grant the AI provider any ownership, licence or other rights over Foundation-funded outputs or the data used to produce them. 

Declaring AI use 

We will ask applicants and grant holders to declare how generative AI has been used. We may document these responses as part of our application, due diligence, monitoring or reporting records. 

This may include asking: 

• Which AI tools were used 

• What the tools were used for 

• What content or activity the tools supported 

• How the outputs were reviewed or checked 

• Whether any personal data, confidential information or unpublished intellectual property was used 

• Whether AI generated content forms part of any output 

We may request access to records of AI use at any time during the grant period and for a period of six years after the grant ends. Grant holders must retain sufficient records to demonstrate compliance with this Policy. 

Keeping this Policy up to date 

We may update this Policy from time to time at our  sole discretion. 

It is the responsibility of applicants and grant holders to review the guidance periodically and ensure they are familiar with the most recent version. The version published on the Foundation’s website shall be the authoritative version at any given time.