Love it or hate it, ChatGPT is here to stay

No doubt you’ve been hearing about, if not using, ChatGPT since its launch last November. It certainly landed with a bang – within a week it had one million users, and by January it had crossed the 100 million users milestone and was seeing 13 million active users a day.

The release of the AI-powered chatbot (and countless emerging alternatives) has sparked intrigue and debate, as the world comes to terms with the consequences of generative AI across everything from education to tech. In fact, the International Baccalaureate is now allowing students to quote from content created by ChatGPT, and programmers are using the chatbot to solve obscure coding challenges in seconds.

Love it or hate it, ChatGPT is here to stay. So, what are some real-world applications of the chatbot for comms and behaviour change, and how has AI already contributed to meaningful behaviour change?

ChatGPT for behaviour change comms

Arguably the most obvious use case for ChatGPT in comms (and generally) is research, with questions already being raised as to whether Google’s reign will soon be over.* No need to sift through pages of results, and the opportunity to ask for results in layman’s or academic terms. You can also brainstorm with the chatbot; one user reported asking it to roleplay as a client and consider a potential campaign – ChatGPT, in its long list of recommendations, suggested the user first take cost and logistics into account and consider how the initiative could fit into a larger integrated launch campaign.

And when it comes to writing, the chatbot can be invaluable in helping you to get ideas down on paper, or to finetune existing ideas. Paste your article into the chat and ask for a few potential headlines, or alternatively give it a headline and receive advice on structuring an unwritten piece. It can even generate compelling fundraising copy from an article. It might take some back and forth – and don’t be afraid to ask it to change its output to serve a specific purpose and increase engagement – but it’s undoubtedly a time-saver.

ChatGPT can speed up processes when you’re faced with a behaviour change challenge, too. With its TL;DR (too long; didn’t read) function, it can cut the literature review process in half by returning a summary of key points. It can also code and analyse open-ended survey responses or focus group content, identifying repeated themes and categorising accordingly. Perhaps even more impressively, it can carry out COM-B analyses and provide evidence to back up its evaluations, and devise behavioural science strategies and interventions. You can ask ChatGPT to use existing, and create new, behavioural science frameworks too; the chatbot has written OASIS plans, and come up with new, well-designed frameworks with nice acronyms. Step aside, humans… right?

Well, the chatbot isn’t without its flaws as it exists currently. Beyond the ethical considerations associated with it (there are some unknowns around data privacy so it’s recommended you use fake names, for example where NDAs are involved) its knowledge base goes no further than 2021, which makes it ignorant of recent events. As well as that, it has at times been guilty of producing incoherent ‘gibberish’, providing fake (but realistic) quotes from real people, and in one instance, producing a fully coherent explanation of a non-existent phenomenon.

But even aside from the tech being in its infancy, just like a calculator that helps a user arrive at a solution, it is thought that AI-powered chatbots should only be considered instruments which can help us reach our goals. Quality in, quality out, as they say – so maybe AI won’t replace us yet, but the people using AI might.

AI-powered behaviour change initiatives

Although ChatGPT is just emerging, AI has been central to innovations designed to drive social change for some time. Mental health apps are an excellent example of this: in a world where mental health resources are scarce and expensive, therapy bots are increasingly filling the gap. No waiting lists, sessions are available around the clock, and they’re cheap (if not free). Plus, research suggests that people may be more comfortable opening up to a bot than a human.

The ‘emotionally intelligent’ AI Wysa is currently being rolled out to state schools across West London and trialled in the NHS to explore how it could help millions sitting on waiting lists. More research is needed to establish the extent of therapy bots’ potential; but with progress happening fast, the hope is that they will soon be a source of relief alongside a better-functioning mental healthcare system.

AI is contributing to advancements in physical health, too. AI-powered personal trainers have become a viable alternative to face-to-face coaching, offering tailored recommendations for workouts, diet, and lifestyle, updated in real-time and using comparison data from millions of others. This integration is making health and fitness more affordable, accessible and engaging for many.

And personalisation is reaching new heights this year, with technologies expanding to analyse user DNA to inform recommendations, minimising the risk of exercise-related injuries for those who are genetically predisposed.

So, what does this all mean for AI’s role in behaviour change in the future?

Ultimately, whether AI will be a force for good – or bad – remains to be seen, but its potential for contribution to meaningful social and behaviour change is unmistakeable.

*Since writing this blog, Google announced a number of AI features coming to Google Docs and Gmail.

Claremont will be hosting a free online event on 30th March: “Can AI Help Change Behaviour for Social Good?” Sign up here.