Learning from experience: How to use AI and legaltech intelligently

Freeths is at the forefront of integrating artificial intelligence (AI) and legal technology (LegalTech) into its practice, including an extensive trial of Microsoft 365 co-pilot and being the first UK law firm to adopt Document Drafter. Freeths experts on AI and LegalTech, Will Richmond-Coggan and David Lane, share the lessons they’ve learned so you can make the most of the opportunities while minimising the risks. 

1. Get the ball rolling in the right way
   
   Establish a secure enterprise-level solution for allowing people access to AI at work as, if you don't, employees will find their own way which brings a number of risks. We advise restricting access to any publicly available AI tools which could pose a threat to your data confidentiality; providing access to a secured and vetted AI solution; establishing a clear policy framework for when and how that AI solution can be used; and empowering staff to learn how to use it responsibly and safely.
   
   
2. Let people experiment
   
   Once you have a secure system in place, make time for people to get together and explore what the technology can do. Freeths organised AI hackathons that were collaborative, creative and educational, to help participants understand and leverage AI technology in their work.  
   
   
3. Treat AI as an enthusiastic assistant
   
   AI will get things wrong but, if you work with it, you can be more effective than working on your own. Look for tasks like drafting or brainstorming where a human can review, build on and enhance the tech’s contribution. Use AI as a general productivity tool and focus on low hanging fruit such as rewriting emails or summarising documents. 
   
   
4. You can’t do it all
   
   Don't let the range of options overwhelm you. LegalTech and AI tools can touch almost every part of your business, but we found it's best to focus on one way in which AI can help, show colleagues the value in that and grow experience in the business so people are ready to try AI themselves.
    
5. Building AI agents takes time
   
   Building your own AI solutions is a very significant commitment. Don't underestimate the technical challenges associated with implementing agents including data quality and hygiene issues, and the complexity of integrating AI systems within existing workflows. Identify a limited and achievable goal; develop the solution then learn lessons from that. 
   
   
6. Proceed with a degree of caution
   
   The regulatory position in the UK is far from clear at the moment; the AI Opportunities Action Plan is long on ambition but short on detail. Without clear regulation and a regulator in the UK, we expect to see AI being used up to the point at which it produces a harm and litigation following. That litigation will be long and costly, will damage reputations and drain resources, so think carefully about the implications of using new technology.
    
7. Go above and beyond to assess impact 
   
   If you're looking at deploying an AI tool or if you're building and developing AI-powered products, carry out an impact assessment. Your assessment should look beyond privacy and use of personal data to risks of bias, the security of any data, the integrity of the systems and societal impacts of the technology including environmental impact. Make sure that you're deploying AI in a way that's consistent with your own policies – BCorp status for example – and document what you do.  
    
8. Think about how risk is allocated
   
   If you're buying in a tool, ask what warranties, assurances or indemnity against harms the supplier will give you. If you’re using that technology to interact with employees or customers, think about what you include in terms of use. For example, if you've got a chatbot, be clear about what it can and can’t do and that anything it says isn’t a legally binding commitment on behalf of the business.