AI is already being applied by behaviour change practitioners, creatives, comms professionals and campaigners.
In this blog, we explore its transformative potential, introduce some basic terminology and suggest a series of tools and use cases to start experimenting with.
There is a caveat before we begin. The tools that we introduce in this blog are great in the early phases of a project, but the results they produce will not be good enough (yet) to base behaviour change interventions or evidence-based campaigns on. However, the ability to create and model behaviour using believable proxies of human behaviours looks to be coming in the very near future according to a new paper from Stamford University and Google Research.
For now, think of AI agents as time and energy savers and always have a human sense check in the process.
AI for beginners
There are some important terms to get used to. You’ll have heard people talking about AI, machine learning and generative AI.
But what are they and how are they different?
AI is a branch of computer science that deals with the creation of intelligence agents, which are systems that can reason, and learn, and act autonomously. Essentially, AI has to do with the theory and methods to build machines that think and act like humans.
Machine learning is a sub-discipline of AI. It is a program or system that trains a model from input data. That trained model can make useful predictions from new or never before seen data drawn from the same one used to train the model.
Finally, Generative AI is a type of artificial intelligence technology that can produce various types of content, including text, imagery, audio, and synthetic data. It creates new content based on what it has learned from existing content.
I work in communications and behaviour change. What can AI help me do?
There is an almost in-exhaustible number of use cases for AI in behavioural science and communications. Here are some examples of where we see its true value:
- A/B Testing: AI-driven A/B testing tools can optimise communication materials by analysing which versions resonate better with the audience. This can lead to more effective messaging.
- Chatbots: AI-powered chatbots can provide instant responses to customer inquiries, enhancing user experience and collecting valuable data on user interactions.
- Content Generation: AI-driven content generators can create engaging blog posts, articles, and social media content, saving time and ensuring consistency in messaging. While this can be useful it is very important to remember that AI is unlikely to account for localised language, terminology or phrasing that your target audience uses. It’s important to use the right prompt for content intended for a place-specific audience and to edit the content generated.
- Data analysis. Uploading files and prompting analysis and basic data visualisation.
- Email Marketing: AI can personalise email marketing campaigns based on recipient behaviour, increasing open rates and conversion rates.
- Insight and evidence gathering. Recommending previous studies and examples of work.
- Language Translation: AI-powered translation tools can translate content into multiple languages, but always check the results with a fluent person.
- Sentiment Analysis: AI tools can analyse social media, service reviews (e.g. council or NHS services), and feedback to gauge population sentiment and identify trends. This insight can be invaluable for crafting targeted communication strategies or service improvements.
- Social Media Management: AI tools can schedule posts, analyse engagement metrics, and recommend content strategies for social media platforms, helping organisations maintain a strong online presence.
- Speech Recognition: AI-driven speech recognition software can transcribe spoken words into text, making it easier to analyse and disseminate important information.
- Population Segmentation: AI can segment populations into distinct groups based on their behaviour and preferences, allowing organisations to personalise their communication and other interventions.
- Predictive Analytics: AI-powered predictive analytics models can anticipate people’s behaviour based on historical data. This can help test behaviour change intervention design.
There are lots of inexpensive tools to get started. Many of them are free to sign up for.
We strongly recommend that teams play with AI as a way of de-mystifying it and finding ways of using it to help or speed up their work. Once you start playing with the tools you’ll soon discover the importance of a well written prompt (the request you make to the AI). Of course, you can also ask the AI to write you a good prompt for the task that you want to complete. Creating better prompts is a topic that we explore during our AI hack events.
Let us know if you’d like our help organising a team AI hack event.
Getting started with AI – from the Magpie Toolkit
Learning resources
- Google’s Free Course on Generative AI is a good introduction for anyone.
- IBM has committed to train 2 million in Artificial Intelligence in three years, with a focus on underrepresented communities through the free IBM SkillsBuild education program.
Have a play
- ChatGPT. Probably the one you’ve heard about and the easiest one to start playing with. Just sign up for free and ask it to do something.
- Bard: An experiment by Google that lets people in your organisation collaborate with generative AI. Your organisation might need to enable access to Bard.
- Claude. We especially like Claude as a great time saver if you have lots of documents to read. Upload a PDF and ask for a summary. Great for contracts, technical reports and published articles. Don’t rely on it for literature reviews though.
- Bing Chat. The use cases write themselves. Especially useful for organisations whose teams aren’t on Google Workplace.
Save yourself some time. Our top five favourite AI tools at Magpie (December 2023).
- Otter.ai AI meeting assistant that records audio, writes verbatim notes, captures action items, and generates summaries. Very useful for focus groups and interviews where transcription would previously have been needed. Another tool, Tl;dv offers similar functionality to Otter.ai.
- Copy.ai is described as a ‘brainstorming buddy, blank page remover, research assistant, and expert copywriter’. It can speed up almost any copywriting task.
- Dialogflow is a tool for creating lifelike conversational AI for chatbots and voicebots. Lots of opportunities here for public services that are under extreme pressure.
- Chatbase. Makes it easy to build a chatbot for your own website.
- Perplexity. Built for people wanting to find the source of information as well as a summary of it. Good for information discovery and curiosity.
Want to discover more or talk to us about hosting an AI hack with your team? We’d love to talk.