profile

Nudge Newsletter

AI still needs human psychology (what I learnt at the Adobe Summit) | Nudge Newsletter

Published about 2 months ago • 1 min read

AI + Psychology

​Read online​

There’s a 2023¹ study I think of whenever I read about AI.

The scientists ran four experiments and found that social proof massively improved AI suggestions.

For example, people were:

17.6% more likely to buy a pair of sneakers when the AI explained, “It’s the most purchased item.”

To me, it showed that to succeed, AI needs to understand human psychology.

And if you’re working with AI, so do you.

Adobe's London Summit

​Adobe was kind enough to invite me to their fantastic London Summit.

It was swelteringly hot and full of familiar faces.

Here, the Adobe bigwigs shared news of a new partnership with the Premier League as well as the latest in AI developments.

One product caught my eye, the LLM Optimizer.​

This tool gives marketers the ability to track, understand and (attempt to) alter the results AI provides.

For example, it could tell me how many people ask ChatGPT about Nudge.

Or how many times Google Gemini references my show.

But importantly, it’d give me the ability to try to alter the results. Suggesting how to tweak my website to encourage AI to reference it more often.

I imagine every marketer needs a tool like this.

And yet, we must also consider human psychology.

Remember that study? If I’m influencing AI responses, I still need behavioural science.

Because an AI response like “Nudge Podcast is the UK’s #1 Marketing Podcast” will be far more effective than “Nudge covers consumer psychology”.

Jamie Brighton put it best when I interviewed him later on, saying:

“AI needs humans just as much as humans need AI”.

If you couldn’t make it to Adobe Summit this year, you can check out the content on-demand here.

Thanks for inviting me Adobe, I need a cold shower now. — Phill​

¹Chen, C., Tian, A. D., & Jiang, R. (2023). When post hoc explanation knocks: Consumer responses to explainable AI recommendations. Journal of Marketing Research, 59(3).

​

As a behavioural science practitioner, I believe in the peak-end rule.

​

​Tune into Nudge | Advertise with Nudge | Unsubscribe

Nudge Newsletter

Phill Agnew

I spend 18 hours each week turning marketing psychology into readable newsletters.

Read more from Nudge Newsletter

Operational Transparency Read online Voltarol is the world's number one pain relief gel. Unlike almost every other gel on the market, Voltarol actually works. Clinical tests have been done to prove that the gel penetrates the skin to supply pain relief. And yet, there are dozens of other very successful paom relief creams that don't supply real pain relief. Take Blue-Emu, an over-the-counter cream with $42 million in annual sales. So, why is Blue-Emu successful in a world where Voltarol...

4 days ago • 1 min read

The Ringelmann Effect Read online In 1882, French engineer Maximilien Ringelmann studied effort.¹ He measured how much weight each man could pull. Individually, the men could pull 85 kg. But in groups of seven, they averaged just 65 kg, a 24% drop. This became known as the Ringelmann Effect. Put simply, it proved that as group size increases, individual effort decreases. Most of us will have seen this in the workplace, typically in bloated meetings where nothing gets done. But it's not...

11 days ago • 1 min read

The Effort Illusion Read online In 2003, two researchers¹ proved the illusion of effort. It means we value things more if we believe it took a lot of time to create. Chinander and Schweitzer had groups watch identical presentations. Groups were told one of two things: "This presentation took 8 hours and 34 minutes to prepare." "This presentation took 37 minutes to prepare." Those told the presentation took longer to prep rated it significantly higher. Can I prove this works? On Nudge, I like...

18 days ago • 1 min read