Take The AICSA Responsible AI Pledge
We invite individuals and organisations to take the AICSA Responsible AI Pledge and make a clear commitment to the ethical, safe and transparent use of artificial intelligence.
Together, these commitments can help build greater trust in AI and ensure it benefits people, organisations and wider society.
Take The Pledge – Individuals
- Human Accountability
I will stay accountable for the decisions I make with AI. The tool is not a place to put the blame.
- Verify Before I Use
I will check AI generated content for accuracy before I share it, publish it or act on it.
- Honest About My Use of AI
I will be open about where and how I use AI, and I will not present AI generated work as something it is not.
- Protect the Data
I will not put confidential, personal or sensitive information into AI tools that have not been approved to receive it.
- Treat AI as Attack Surface
I will treat the AI systems I use as part of my attack surface and follow the security controls that apply to them.
- Challenge Bias
I will look for bias in AI outputs and raise it rather than pass it on.
- Respect Consent and Credit
I will respect intellectual property, consent and attribution, and I will not use AI to take someone else’s work and call it my own.
- No Harm By Design
I will not use AI to deceive, impersonate, harass, manipulate or exploit anyone.
- Privacy By Default
I will handle personal data lawfully and use the least amount needed for the task in hand.
- The Right Tool, Not the Fastest
I will ask whether AI is the right tool for the task before I reach for it.
- Judgement Over Shortcuts
I will keep building my own knowledge and skills so that my judgement stays sharper than the tool I am using.
- Human Review Where It Counts
I will make sure a person reviews AI outputs before they affect anyone’s safety, rights, livelihood or wellbeing.
- Report Failures and Near Misses
I will report AI errors, harms and near misses rather than quietly working around them.
- Explain, Do Not Exclude
I will explain AI outputs and decisions in plain language, avoiding jargon that shuts people out.
- Stand With Those Who Speak Up
I will support people who raise concerns about the use of AI and I will not leave them to face it alone.
Take The Pledge – Organisations
- Named Accountability
We will name a senior person who is accountable for how AI is used across our organisation, and we will make it clear to our people who that is.
- A Policy People Can Actually Follow
We will maintain an acceptable use policy for AI that is written in plain language, kept current and easy to find.
- Know What We Are Running
We will keep an inventory of the AI systems we use, including AI features embedded inside third party products and services.
- Assess Before We Deploy
We will assess the safety, security, privacy, bias and legal implications of an AI system before we put it into use, and we will record what we decided and why.
- Secure the AI Supply Chain
We will apply due diligence to the models, data, plugins, APIs and vendors we depend on, and we will not assume a supplier has done it for us.
- Test, Monitor and Test Again
We will test our AI systems for misuse, manipulation and failure before launch, and we will monitor them for drift and unexpected behaviour afterwards.
- Protect Personal Data
We will process personal data lawfully, minimise what we collect and complete a data protection impact assessment where one is required.
- Human Oversight Where It Counts
We will not allow AI alone to make decisions that materially affect a person’s rights, safety, livelihood or access to services, and we will give people a clear route to question a decision.
- Be Transparent With People
We will tell our customers, employees and partners when they are interacting with AI or when AI has played a meaningful part in a decision about them.
- Train Our People
We will give our people practical training on using AI safely and securely, and we will refresh it as the technology and the threat landscape change.
- Safe to Raise Concerns
We will make it safe for anyone to raise a concern about our use of AI, and we will protect those who do.
- Plan for Failure
We will include AI failures, misuse and compromise in our incident response planning, and we will rehearse them.
- Review and Improve
We will review our AI governance at least once a year, and sooner if the technology, the regulation or our risk profile changes.
The 3 Actions in Your First 90 Days
Principles matter, but dates are what move them from intention into practice. Organisations signing the pledge commit to complete the following 3 actions to the timescales shown, measured from the date you sign.
Action 1: Make AI a Board Responsibility
- Name a board level or senior executive owner who is accountable for how AI is used across the organisation, within 1 month of signing.
- Put AI risk and AI governance on the board or senior leadership agenda within 3 months of signing, and at least annually after that.
- Ensure everyone at board or senior leadership level completes AI governance and AI risk familiarisation within 3 months of signing, and refreshes it annually.
Action 2: Know What You Are Running
- Begin an inventory of the AI systems in use across the organisation within 1 month of signing, including AI features embedded inside third party products, browser extensions and tools bought on expenses.
- Complete the first full inventory within 3 months, recording an owner, a purpose and a risk rating for every entry, and present it to the board or senior leadership.
- Review and refresh the inventory at least every 6 months, because shadow AI arrives faster than policy does.
Action 3: Set A Responsible AI baseline across your supply chain
- Publish or refresh an acceptable use policy for AI, written in plain language and accessible to every employee, within 2 months of signing.
- Add responsible AI questions to your procurement and supplier due diligence process within 3 months of signing, covering how suppliers use AI, on what data and with what human oversight.
- Take a risk based approach to the assurance you require from suppliers. Where you decide a lighter touch is appropriate, record the reasoning and make sure it sits within your risk appetite.
What Else Can Organisations Commit To?
Alongside the 3 actions, organisations signing the pledge also commit to the following.
Encourage these actions within your own supply chain. Engage with your suppliers to understand how they are using AI and encourage them to adopt the same measures.
Publish your signed declaration on your website within 2 months of signing, so the commitment is visible to your customers, your people and your partners.
Publish an annual public update on the steps you have taken to deliver against the pledge, either in your annual report or on your website.
Tell us if you cannot deliver. If you have been unable to implement any part of the pledge, let the AICSA know why. That feedback shapes how the pledge develops and it is far more useful to us than silence.
Staying Signed Up
The pledge runs on a 12 month cycle. At the end of each cycle we invite signatory organisations to reaffirm their commitment and to undertake the actions again for the year ahead, because the threat landscape, the technology and the regulation will all have moved.
The AICSA will review the pledge itself at least every 12 months and may refine the actions as the field develops. Where we do, signatories will be told what has changed before they are asked to reaffirm.
Signing is voluntary and so is staying. You can withdraw at any time by emailing us and we will remove you from the list. Equally, where an organisation is clearly not acting in the spirit of the pledge, we reserve the right to remove it.