Skip to main content

Behavioural AI: What Is It? And How Can We Use It?

Dec 20 2023 • 4 min read

If you work for a large organisation anywhere in the world, the chances are that you will have been asked at some point: “what can we do with Generative AI”?

Generative AI systems, most notably ChatGPT, have the ability to “generate” high quality content – in the form of text, images, code, or music – in response to prompts from their human users.

The advantage that the latest wave of AI systems have is that they have been trained on vast quantities of data (i.e. more or less the whole internet), which is then used to learn how to complete a task in response to a prompt.

So for some people, the answer to the question of what to do with Generative AI is obvious. You can prompt a Generative AI system to create the code and designs for a new website, generate the content for your first set of articles, and produce photo-realistic images to help bring them to life.

Through these endeavours, the goal of many AI researchers has been to make machines capable of completing tasks that previously only human beings were capable of performing. And to do so at a speed and accuracy that would have previously been impossible to match.

This has led many commentators to consider whether this might herald the "end of the human era" or even the extinction of the human species. But while we firmly believe that we should be cognizant of the nefarious potential of uncontrolled AI, we also believe it would be wrong to suggest that human beings and machines are currently competing in a singular Darwinian struggle for survival.

At CogCo, we believe that we need to focus on the potential symbiotic opportunities between human and artificial intelligence. We are currently doing this in two different ways, both of which use AI to support and enhance the application of behavioural science and design methods. We are calling them ‘Behavioural AI’.

The first (relatively simple) application involves using AI platforms to process and analyse huge quantities of data; and using behavioural science and design knowledge to inform the analysis and nudge actual behaviour. Existing platforms have already been trained on publicly available methods, papers and frameworks, but it requires behavioural and design expertise to prompt and refine these systems in the right way. And further expertise to understand what additional inputs and insights (such as proprietary frameworks or privately held data) might be required to enhance the outputs generated.

The second (relatively more complex) application focuses on using our expertise in the understanding of human behaviour to train existing AI models to ‘think’ behaviourally. What it means in practice is our partners having access to a model that can automatically recognise different barriers and drivers, and produce solutions that change people’s behaviour in the real world.

Imagine, for example, that we want to help teams of people in a call centre to improve how they engage with customers (demonstrated through some change in a metric such as Customer Satisfaction Score). We might ordinarily gather these insights through a combination of listening in on a sample of calls; conducting qualitative interviews; or running a large scale survey.

But with the latest generation of AI systems, we can use the first of our Behavioural AI applications to process every single call made in the past year; transcribe these into text; and then use specific behavioural and design frameworks (e.g. the Theoretical Domains Framework or CogCo’s behavioural diagnostic methodology) to prompt an AI model to draw out the most significant behavioural factors associated with high Customer Satisfaction Scores. We can then use behavioural experts to use these insights to generate solutions that are designed to change human beings’ behaviours.

We can take this a step further through the second application of Behavioural AI by training existing models to automatically identify behavioural drivers and barriers in calls coming into a call centre; and to immediately generate prompts and solutions that can be actioned by human beings in the real world.

CogCo’s Behavioural AI practice is currently running pilots in these areas. So if you’re one of those individuals who is currently musing on the question of how to use AI, but wants to bring the human back into the frame, drop us a line.

Read recommended blogs

Human, informed by Computer, Says Yes Human, informed by Computer, Says Yes

In the comedy show Little Britain there is a sketch involving a customer service representative sitting behind a computer. Whenever a customer makes a perfectly reasonable request, she taps away at a keyboard and says: 'computer says no'. The sketch resonated with viewers because it illustrated what can happen if you do not engage and...

Feb 7 2024 • 3 min read

Feb 7 2024 • 3 min read

Artificial Intelligence, Chatbots and Behavioural Insights Artificial Intelligence, Chatbots and Behavioural Insights

If you’ve ever tried to contact a bank with a query about an account, a utility provider over how to make a payment, or a retailer about a missing delivery, the chances are that you will have been forced to engage with a chatbot. And the likelihood is that, with a few exceptions, you will...

Jan 12 2024 • 4 min read

Jan 12 2024 • 4 min read

Automatic Enrolment: A New Application in Savings Automatic Enrolment: A New Application in Savings

One of the most cited applications of behavioural science in practice is the powerful effect of automatically enrolling people into schemes like pension plans. With a simple change in the sign-up process, these interventions show that we can move from a world in which the minority of people save for retirement to one in which...

Jan 8 2024 • 3 min read

Jan 8 2024 • 3 min read

Interested in working with us?

Get in touch at

Copied to clipboard