Week 84: The ideological Turing test
This week we had Professor Tom Stafford come and give our Thoughtful Thursday talk. Tom is a friend of mine, and I invited him in after a discussion we had about vaccinations when out walking in the Lake District last year.
Tom discussed 2 of the psychology research project he’s worked on which looked at vaccine hesitancy.
The first was Using dialogues to increase positive attitudes towards COVID-19 vaccines in a vaccine-hesitant UK population Open Access and aimed to replicate a French study which showed that a 5 minute interaction with a chatbot could lead to increases in positive attitudes towards vaccinations. As well using the chatbot style interface (which wasn’t AI based but more like a branching narrative interaction), they also had a control group presenting the same information in a static format. Both groups had the same positive effect.
Their conclusion was that it was engagement that counted, not the interface, and that people’s minds could be changed.
It’s also worth saying though that the study measured attitudes rather than action, so there could still be a motivation-volition gap.
The second study explored the ideological Turing test. This puts a spin on the original test proposed by Alan Turing to distinguish human and artificial intelligence.
The idea is to test whether people are able to summarise an argument which you don’t believe, in such a way that your opponent would agree with. In short: can you at least listen to and understand another point of view?
The project tested this by asking people to give opposing arguments on 3 topics: veganism, COVID-19 vaccines, and Brexit. Their arguments were then shown to the other side, who were asked to rate them.
The results were that, on all 3 topics, people were able express convincing opposing arguments to a reasonable degree. The one exception being veganism, where both pro and anti-vegans struggled to argue the case for eating meat.
You can explore the data, including keywords from the arguments people gave on each topic, in an interactive explorer tool.
Tom’s conclusions from both studies are that people are more rational and willing to engage in complex topics than the media might sometimes portray. Misinformation is not a virus.
In the questions after we also discussed trust. People’s trust in institutions may be reduced, and there’s less deference to doctors than there was. But trust needs to be earned, and perhaps having a more engaged patient population is a good thing, even if it takes more time to explain and reassure?
There’s also a question for us digital folks: should we focus on the ‘low-hanging fruit’ of reducing frictions and making for those who want it to actually get their vaccinations? Or do we have a role to play in decreasing vaccine hesitancy too? I think we can contribute to both, even if the latter is harder.
My thanks to Tom for giving the talk. He’s published the slides online. His Reasonable People substack is worth a follow too.
Links
- What is user research? on YouTube is a fantastic video from my colleagues in the breast screening team, where Fred manages to show (not tell) the value of their ongoing user research.
- Personalised Prevention Strategy – an update in 2 stickers from Irina and Naomi gives an update on the strategy from their team.
- Children to be better protected from second-hand smoking and vaping from the Department for Health and Social care kicks off a 10 week consultation on the extension of the smoking ban to include outsides spaces like playgrounds and hospital grounds. I do wish that this included pub beer gardens too though.
I just had a jacket potato for the first time in ages. So good, and an antidote to this miserable weather.