This article was written by Deelan Maru, Elisabeth Costa and Michael Hallsworth from the The Behavioural Insights Team.

AI promises to transform the way public servants work, but to do so it needs to be meaningfully used. That is an inherently behavioural challenge. Thoughtful adoption is not about replacing human skills, but augmenting them - working in partnership with AI to achieve more and freeing us up to focus on tasks that require our judgement and experience.

Many of us already use AI tools in our day-to-day work. But what if we’re taking the small wins and leaving the bigger gains on the table? Behavioural science calls this ‘satisficing’. We use tools to take meeting notes, summarise reports or draft correspondence - all worthwhile use cases that should continue. Yet AI could be used in much more value-enhancing ways: improving enrolment across services; using simulated scenarios for developing policy; or assisting caseworkers to triage cases more quickly.

So, what’s holding us back?

In our recent report we argue that AI adoption is at its heart a behavioural challenge. How do we shift our day-to-day work habits to integrate AI in thoughtful, value-adding ways? There are three key barriers and enablers:

  • Motivation: do you see a clear, desirable reason to use AI? AI discourse can frame benefits around ‘increased productivity’ or ‘improved decision-making’, which can be disconnected from an employee's actual tasks.
  • Capability: do you feel able to use it effectively and confidently? Estimates suggest that c.70% of adoption challenges stem from people and process issues rather than technical constraints.
  • Trust: does AI align with your values? Does it present a threat to your identity? In one study, researchers observed a social penalty for using AI, where people who used it were consistently rated as lazier, less competent and less diligent.

How, then, can you identify and remove these behavioural barriers?

Firstly, diagnose issues within your teams. You can use surveys, behavioural systems mapping and sludge audits to do this. Once you’ve identified the barriers, the next step is to design solutions to overcome those barriers. The most impactful results will come from genuine co-design with your teams, but here are three example interventions you may wish to try:

  • Inspire with examples. Team members may simply not know what kinds of deeper uses of AI are possible and many others may be using AI in ways that enhance their work, but not telling anyone about it. Actively demonstrate what's possible through curating and sharing a library of role-specific use cases.
  • Use social proof. Make AI use visible and celebrated to reduce identity threats. Those who know someone who has used AI are three times more likely to have used AI themselves. This so-called ‘bandwagon effect’ can drive adoption at speed and scale, with minimal effort.
  • Evaluate impact and embrace the results (positive or negative). Knowing what works, particularly with high-quality evidence to back it up, can strengthen deeper adoption. Alongside this, celebrate null results. Acknowledge what didn’t work, and why it didn’t work. This creates psychological safety for further experimentation, as your team won’t fear retribution if their idea doesn’t work.

For any solution you develop, take a test & learn approach, utilising rapid prototyping and iterative feedback cycles to quickly understand what works and pivot as needed to drive adoption. The best teams will use each experiment as a way to learn, treating AI adoption as an ongoing endeavour.

__Adopt is one part of a new BIT framework exploring four fundamental issues facing AI. The full AI & Human Behaviour series is available to download now. __

At BIT, we help organisations assess and expand their use of AI in thoughtful ways through training - like our course Accelerate AI adoption with behavioural science - and projects that uncover and address the real barriers to AI adoption.

Get in touch [deelan.maru@bi.team] to discuss how we can work with you.

Make sure to share your own thoughts with the author by leaving a comment below