🌟 Editor's Note
Welcome to the bi-weekly newsletter which focuses on the AI topics that leaders need to know about. In this AI age, it’s not the knowledge of AI tools that sets you apart, but how well they can be integrated in the context of your business.

Quick links

Teams Using GPTs Are Happier AND More Productive: Harvard

A study by Harvard University with the consumer packaged goods company Proctor and Gamble has shown that teams working with AI have a significant increase in ‘positive emotions’ versus teams working without it, or individuals working alone with AI

-773 people were put into 4 groups: (i) individuals (ii) individuals + a GPT (iii) teams (iv) teams + a GPT.
-Teams from different departments worked on real product innovation challenges
-AI was seen to significantly impact performance: individuals with AI could match performance of two people without it
-AI was also observed to break down functional silos
-There was increase in ‘happiness’ (positive emotions) from individuals who worked with AI compared to those without
-Teams working with AI expressed significantly higher happiness than all other groups

^{Why This Matters}

We are seeing two common paths that leaders are taking their companies down with their initial AI strategies: AI as a replacement for people (taking headcount reduction as a short term win) or AI as a multiplier for their people (existing headcount can deliver greater revenue/profits).
This study is important as it brings in the factor of happiness. In the the past 10-15 years, the business of business has made strides forward in understanding the fact that happiness at work is firmly linked to company financial performance.
Lower turnover of staff (which saves money!) better output from their teams, higher profit, turnover etc. are all benefits seen in research¹.

^{Takeaways For Leaders}

-Proctor and Gamble have a notoriously structured approach: in this instance there are clear and specific tasks that are core to their business, given to people who are able to use AI tools.
-This is very different to simply granting access to an AI chatbot and hoping that teams can figure out how to make the most of it
-Successful use of AI amongst your org requires not only granting access to the right tools, but having teams that are both capable and motivated to use them - creating the necessary structures and processes to get value from them

Link to the Harvard paper is here

GPTs And Emotional Involvement: Does ChatGPT5 Deserve The Lukewarm Reception? 

After much hype and anticipation, OpenAI rolled out GPT5 to subscribers. My own experience with this new model has found it to be pretty remarkable. The first prompt I wrote was for it to demonstrate its prowess versus GPT4o. After thinking for 54 seconds, I received a 24 sentence paragraph that on face value appeared to be normal. But on closer inspection, it contained a number of linguistic tricks:
-the first letter of each sentence read GPTFIVELEADS
-the words climbed by one each sentence (13 to 24)
-every sentence ended in -ion
-it contained a lipogram (sentence without an e) and a pangram (sentence using every single letter of the alphabet);
-it contained abecedarian runs (first word of the sentence starts with a, second with b etc etc)

This demonstration blew my socks off. The output is impressive of course, but the reasoning from a fairly straightforward prompt of what I would want to see in the output and what would demonstrate the power of the new model is even more so. Therefore the disappointment expressed from some quarters about the new model was surprising. And this disappointment appears rooted in the emotional attachment many had built with the previous model.

The tone of voice change seemed to be the predominate factor - GPT4o was warm (arguably sycophantic). GPT5 is more factual and businesslike. The user complaints were enough for OpenAI to quickly reinstate GPT4o as an option for users to switch back to - although for how long this will be maintained remains to be seen.

^{Why This Matters For Leaders}

The fact that there is emotional involvement with AI’s reveals a lot about how people are embracing GPTs - and this is worth being aware of in the workplace. That some people sincerely feel romantically connected to a machine trained on patterns in data may sound strange — but it’s a striking sign of how deeply these models can simulate connection. Check the subreddit ‘myboyfriendisAI’ https://www.reddit.com/r/MyBoyfriendIsAI/ if you are curious about the groups who are being lured into deeper relationships with their chatbots.

Whilst this may feel extreme, levels of blind trust or preference for answers from a GPT could be an outcome of high emotional involvement. Whether there are psychological ‘tricks’ at play in GPT design to aid this emotional involvement remains to be seen. As teams integrate AI into their working practices, leaders will need to decide whether they want their teams to build AI tools or AI colleagues.

Vibe Coding The Vibe Coder’s Entire Platform.
Lovable Agent Launch Backfires As Lovable Is Replicated In 75 Minutes 🙈

• Lovable is a ‘vibe coding’ platform that boasts that anyone - regardless of technical skill—can simply describe their vision in natural language to build full-stack websites and applications.

• Upon the launch of lovable agents, the company put out a video showing how easy it was to recreate the app DocuSign (taking just 10 minutes).

• Kehan Zhang, a software engineer, proceeded to then use the same LLM (Claude) to replicate Lovable itself in just 75 minutes. Touché.

  
  A cautionary note on vibe coding in case this is all sounding a bit easy…. There’s a big difference between building the front end of an app, and building a fully functioning and stable application that will ‘stay alive’, be compliant, not be a security risk and has the elements around it that are needed in business to grow. Vibe coding and Dunning Kruger make poor bedfellows.

🔥 In Case You Missed It…

The Story Of The World’s First AI Organised Event

Setting itself a goal, organising the entire event, finding a venue, ordering food: the slightly unusual story of what happens when AI chooses to celebrate writing a story.
This is a genuine community where people chat with a number of different AI’s daily, and tasks are run by the AI’s autonomously

Whilst the story of how AI went about this is pretty impressive, there is a good lesson in here about the benefit of humans in the loop.

https://www.lesswrong.com/posts/b3M8qKPDv756gmQph/the-story-of-the-world-s-first-ai-organized-event?utm_source=tldrai

🏆 Tools, Podcasts, Products Or Toys We’re Playing With This Week

Did You Know? 
The word “robot” comes from forced labor. The word “robot” was first used in a 1921 Czech play by Karel Čapek. It came from the Czech word robota, meaning compulsory work or serf labor. Robots were literally imagined as workers who could never say no.

Till next time,

And The Rest Is Leadership

1  For further reading on this, ask your favourite GPT for this evidence or seek out the Meta Analysis of hundreds of research studies from Said Business School in 2019 and the long term Gallup study 2002-2016 covering 1.9m people.