Alex is Sprintlaw's co-founder and principal lawyer. Alex previously worked at a top-tier firm as a lawyer specialising in technology and media contracts, and founded a digital agency which he sold in 2015.
Yes, if your staff are using AI at work, you should have an AI use policy. It doesn't need to be long, but it does need to be clear about what staff can put into AI tools, what they need to check, and who owns the work that comes out of it.
The slightly trickier part is that AI use cuts across several areas of UK business law: data protection, employment, confidentiality, intellectual property and customer contracts. A simple ChatGPT policy for employees can miss some of those moving parts.
The good news is that you can get a practical workplace AI policy in place without slowing your team down. The aim is to give staff safe boundaries so they can use useful tools without accidentally exposing your business.
- Do I Need An AI Policy For My Employees?
- Can Staff Put Customer Data Into ChatGPT?
- Who Owns Work An Employee Creates With AI?
- What Should An AI Policy Template Include?
- Can I Monitor How Staff Use AI Tools?
- Can I Discipline An Employee For Breaching Our AI Policy?
- Should An AI Policy Cover Contractors And Freelancers?
- Key Takeaways
Do I Need An AI Policy For My Employees?
Yes, in most UK workplaces, you should have a written AI policy if staff are using generative AI tools for work tasks. Even if only one or two people are experimenting with AI, a policy helps set expectations before something goes wrong.
More of our clients now mention AI tools in their briefs to us. The practical risks for a business are usually customer data, confidential information, ownership of AI-assisted work, monitoring and staff accountability. Those are exactly the points your policy should deal with.
An AI policy is especially useful if your staff use tools for:
- drafting emails, reports, proposals, social posts or website copy
- summarising customer notes, call transcripts or meeting records
- coding, product design, data analysis or research
- creating images, audio, video, presentations or marketing assets
- screening CVs, managing staff performance or helping with workplace decisions
Your policy should sit alongside your employment contracts, staff handbook, privacy policy, information security policy and confidentiality obligations. If your employment contracts are older, it's worth checking whether they properly cover AI-assisted work, confidential information and business systems.
Templates can be a helpful starting point, but AI use is rarely one-size-fits-all. A marketing agency, a care provider and a SaaS company will all need different rules because they handle different data and face different commercial risks.
Can Staff Put Customer Data Into ChatGPT?
Usually, no - not unless your business has approved the tool, checked the data protection position and told staff exactly what information can be used. Customer data, staff data and supplier data shouldn't be pasted into public AI tools just because it feels convenient.
Under the UK GDPR and the Data Protection Act 2018, you need to think about why personal data is being used, what lawful basis applies, how much data is needed, where it's stored, who it's shared with and how long it is kept. Your privacy policy should also cover key points such as purposes, data retention and sharing.
For example, say Priya runs a 14-person accountancy practice in Manchester. A team member pastes a client's full name, payroll details and director loan notes into an AI tool to draft an email. That may feel like a time-saver, but it could create a data protection and confidentiality problem if the tool's terms allow the provider to store prompts, review them or use them to train models.
A safer employee AI use policy position might be:
- staff can use approved AI tools for general drafting and formatting
- staff must not enter personal data, special category data, payment data or commercially sensitive client information unless approved under your privacy and security rules
- staff should anonymise or remove identifying details where possible
- staff must check outputs before sending them to customers
- higher-risk AI uses need sign-off from a manager or data protection lead
If an AI use involves processing personal data and is likely to result in a high risk to people's rights and freedoms, you must carry out a data protection impact assessment before rolling it out. That's a practical risk assessment, not just a form, and it should be done before the tool is rolled out.
Who Owns Work An Employee Creates With AI?
Usually, your business is the first owner of copyright in work created by an employee in the course of their employment, unless you've agreed otherwise. That rule comes from section 11(2) of the Copyright, Designs and Patents Act 1988.
AI makes this more nuanced because there may be several layers to think about:
- the employee's original input, such as a prompt, brief or edited draft
- the AI-generated output
- any human edits that turn the output into finished work
- the AI provider's terms of use
- third-party material that may have influenced or appeared in the output
The CDPA also has a rule for computer-generated works. Under section 9(3), the author is the person by whom the arrangements necessary for the creation of the work are undertaken. In practice, it can still be unclear how that applies to modern generative AI outputs, especially where there's little human input.
Your AI policy shouldn't rely on the default law alone. It should say that work created, developed or edited by staff using AI for work purposes belongs to your business, subject to the law and any third-party terms. Your employment contracts should also include strong intellectual property clauses.
This matters most for customer deliverables. If you sell logos, code, content, designs, research, training materials or software, you need to know whether you can safely give the customer ownership or a licence. Your policy should require staff to record when AI has materially helped create a deliverable, especially where a client contract asks for originality, exclusivity or human-created work.
What Should An AI Policy Template Include?
A good AI policy template should cover the practical rules your staff need every day, not just broad statements about responsible AI. You want it to be clear enough that someone can make the right call when they're under time pressure.
Here is a useful outline to take to a lawyer and tailor before you use it:
- Purpose And Scope: explain why the policy exists and who it applies to, including employees, directors, contractors, freelancers and agency workers where relevant.
- Approved And Prohibited Tools: list which AI tools can be used, which tools are banned, and who can approve new tools.
- Allowed Uses: set out low-risk uses such as brainstorming, formatting, translation checks, research support or first drafts.
- Restricted Uses: identify tasks that need manager approval, such as customer deliverables, legal or HR documents, code deployment or data analysis.
- Data And Confidentiality Rules: explain what staff must not enter into AI tools, including customer data, staff data, trade secrets and contract terms.
- Accuracy And Human Review: require staff to check outputs for accuracy, bias, outdated information, invented references and brand fit.
- Intellectual Property: deal with ownership, records of AI use, third-party rights and client contract requirements.
- Monitoring And Records: explain what AI usage records your business may review and why.
- Breaches And Consequences: link the policy to your disciplinary process and information security procedures.
- Review Process: say who owns the policy, how it fits your AI governance policy and how often it will be updated as tools and business needs change.
Keep the language practical. For example, instead of saying staff must use AI ethically, say they must not upload client files into unapproved tools and must check every AI-generated customer email before sending it.
Can I Monitor How Staff Use AI Tools?
Yes, you can monitor staff AI use in some situations, but it needs to be lawful, proportionate and clearly explained. Monitoring should protect your business, customers and systems, not become a hidden surveillance programme.
AI monitoring might include reviewing:
- which approved tools staff access through work devices
- usage logs for company AI accounts
- prompts and outputs saved in work systems
- downloads, uploads or integrations involving customer files
- security alerts where a staff member may have entered restricted data
If monitoring involves personal data, the UK GDPR applies. Your staff privacy notice and internal policies should explain what you collect, why you collect it, who can access it, how long you keep it and whether it's shared. You should also keep the monitoring limited to what you actually need.
For higher-risk monitoring, such as detailed prompt review, productivity scoring or automated flags that could affect someone's job, pause before rolling it out. You may need to assess the risks, consider less intrusive options and get tailored advice before using the data for employment decisions.
A practical approach is to separate ordinary system security checks from people management. For instance, checking whether customer files were uploaded to an unapproved tool is different from ranking staff by how much they use AI. Your policy should make that difference clear.
Can I Discipline An Employee For Breaching Our AI Policy?
Yes, you can take disciplinary action for a breach of your AI policy, provided the rule was clear and you follow a fair process. The outcome should fit the seriousness of what happened.
Start with the facts. Keep records of the relevant prompt, output, tool used, data involved, customer impact, system logs and any messages about approval. Then look at your policy, employment contract, staff handbook and disciplinary procedure.
For example, Mike runs a 9-person ecommerce business. His AI policy says staff must not upload customer spreadsheets into public AI tools. A staff member uploads a spreadsheet with 2,000 customer names and delivery addresses to draft a shipping update. Mike should first secure the data position, investigate what happened, speak to the staff member and consider whether any customer or regulator steps are needed. He shouldn't jump straight to dismissal without process.
Under ACAS guidance and general unfair dismissal principles, a fair process usually involves:
- investigating the issue before deciding on an outcome
- telling the employee the allegation clearly
- giving them a chance to respond
- considering evidence and any mitigation
- choosing a proportionate sanction
- offering an appeal where appropriate
As at late 2026, many ordinary unfair dismissal claims still depend on a 2-year qualifying period under the Employment Rights Act 1996, but that's changing. From 1 January 2027, the government has said the unfair dismissal qualifying period will be 6 months. Even where an employee has shorter service, a careful process is still the safer and more professional route.
Should An AI Policy Cover Contractors And Freelancers?
Yes, your AI policy should cover contractors and freelancers if they use AI to do work for your business or your customers. You may need to include the policy in their contract rather than treating it like an internal staff handbook.
This is especially important for intellectual property. Unlike employees, contractors often own the copyright in what they create unless there's a proper written assignment to your business. If a freelancer uses AI to create designs, code, copy or strategy documents, you need the contractor agreement to deal with both human-created and AI-assisted work.
Your contractor terms should cover:
- whether the contractor can use AI at all
- which AI tools are approved
- what data and confidential information must not be entered into AI tools
- who owns the work product and any AI-assisted materials
- whether the contractor must disclose AI use
- what happens if the AI output infringes a third party's rights or breaches a client contract
- security requirements for accounts, devices and file storage
If your customer contracts promise that work will be original, confidential or created only by named people, you need to flow those obligations down to contractors. Otherwise, your business may be responsible to the customer even though the AI use happened outside your team.
This is also a good moment to check your onboarding. Before a freelancer starts, give them the AI rules, confidentiality obligations and approval process in writing. That simple step can avoid a messy ownership or data dispute later.
Key Takeaways
- If staff are using AI at work, a written AI policy helps protect customer data, confidential information, IP and workplace standards.
- Staff shouldn't put customer data, staff data or sensitive business information into public AI tools unless your business has approved the tool and safeguards.
- Employee-created work usually belongs to your business if made in the course of employment, but AI-assisted work still needs clear contract and policy wording.
- You can monitor AI use where it's lawful, proportionate and properly explained, especially for security and compliance purposes.
- You can discipline staff for breaching an AI policy, but the rule should be clear and the process should be fair.
- Contractors and freelancers need AI rules too, because IP ownership and confidentiality work differently outside employment.
If you'd like help with your workplace AI policy, employment contract clauses, contractor AI rules or data protection checks for AI tools, you can reach us at 08081347754 or team@sprintlaw.co.uk for a free, no-obligations chat.
Protect your brand
What intellectual property should you protect?
If a name, logo, design or other creative work matters to the business, check who owns it, what permissions you need and whether clearance or registration is appropriate.








