And The Rest Is Leadership

Helping Leaders Translate AI Into The Context Of Their Organisations .

🌟 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.

Featuring

  • Three Things That Matter

  • In Case You Missed It

  • Tools, podcasts, products or toys we’re currently playing with

GPT’s Rot Your Brain?

With sensationalist headlines such as “ChatGPT making us stupid: MIT” and ‘Digital amnesia’: How AI is dumbing down Aussie kids” , the research paper from MIT made waves. However, the authors of the paper were quick to caution against drawing hard conclusions. As they pointed out, the research is exploratory and needs further nuanced research. Here’s some other considerations for context:

-The study was based on just 54 participants
-The study has yet to be peer reviewed - the academic requirement for reliability
-The application of EEG technology to read brain activity is disputed by some in academic circles

Balanced View for Leaders

There’s evidence that over-reliance on ChatGPT can reduce active cognitive effort in specific tasks—and leaders should be aware. But using AI tools responsibly (e.g., pairing them with human thinking first) can enhance creativity and memory.
The study’s preliminary findings should be seen as a starting point for deeper, broader research—not conclusive proof of harm.

Suggested Thoughts For Leadersons

  1. Educate teams: Ensure AI is used as a collaborator—not a shortcut.

  2. Design workflows: For most teams, human ideation should be combined with AI assistance at this stage, as opposed to relying on AI for any task.

  3. Whether in the workplace or at home, always encourage reflection and memory work after using tools that reduce cognitive involvement.

  4. Stay current: Watch for follow-up peer-reviewed studies to shape policy and training.

This chart from the company Magic Numbers shows the proportion of queries that deliver an AI Overview:

Why It Matters: 
The increasing use of GPTs instead of using search engines is leading companies like Google to include similarly AI based overviews in their responses, probably to stem the flow of search queries being lost.
This is said to be having an impact on how many users are clicking on the search rankings - and this impacts the value of the work companies do to get their businesses ranked higher. The world of Search Engine Optimisation will have to re-evaluate the investments it makes in securing higher rankings. If less people are clicking on the links to a website returned from a search query, the value of that activity will fall.

What to watch out for: with lower trends in traditional search volumes, companies that rely on search as part of their marketing strategy need to focus on other activity to replace this. Discovery of brands via GPTs could be a replacement, but take the words of anyone who claims to have already cracked this area with a pinch of salt. The science of how to get ‘ranked’ on a GPT is in its infancy and will likely change in nature as GPTs work out how to make money from this.  

The original Magic Numbers post can be found here

Chat GPT Launches Agents in GPTs

This is a major move. Agents have been a hot topic in 2025. This upgrade allows the GPT to behave more like an autonomous task completer, not just a chatbot. The agents can:

  • Reason across multiple steps

  • Use tools (APIs, code interpreter, file search, browser, etc.)

  • Work towards goals instead of just answering prompts

  • Maintain context across long workflows (not just single inputs)

It’s part of the new "Custom GPTs" rollout — meaning:

You can build a bespoke assistant (like an intern, data analyst, travel planner, etc.) with special instructions, memory, file access, API calls and now: agentic reasoning

  • Potential Real Use Cases for GPT Agents

    • Exec Assistant GPT: Books flights, prepares meeting notes, fetches documents

    • Research Synthesizer: Takes a question, pulls data from PDFs, summarizes insights

    • Customer Support Agent: Handles inquiries, integrates with live APIs, logs tickets

    • Newsletter Creator: Gathers AI news, formats it, sends drafts to newsletter creation tools such as Substack or Beehiiv (something that this author will be trying to do!)


    The way that agents interact within the web is very different from how we currently use the web. This could herald the way for a very different iteration of the web for the future.

What’s the takeaway for leaders?

Agents offer the promise of automation (and therefore timesaving) across our orgs. HOWEVER ! Before rushing to build agents, leaders need to be cautious that their teams are building in the context of the needs of the organisation AND deploying agents in a safe and secure way. Operating with caution in planning and deploying an agent is wise: they come with security risks, the potential to leak data, the possibility of being compromised by nefarious actors.

Leadership in the AI age requires applying crystalised knowledge and taking educated risks. Agents offer much advantage for an organisation but proper oversight for compliance is essential. This newsletter fears that in the rush to fulfil an ‘AI strategy’ we will see some horror stories within the next 12 months of well intended but badly executed agent implementation.

The words of Alexander Pope are worth taking to heart: “Fools rush in where angels fear to tread”

🔥 In Case You Missed It…

AI Band Passes 1.5m Listeners on Spotify
  • Passing 1.5m monthly listeners, and more than 2m streams of their track ‘Dust on the Wind’, the Velvet sundown have admitted that the music, backstory and images were created by AI

🏆 Tools, podcasts, products or toys we’re playing with this week

Travis, who features in Flesh and Code

Podcast - Flesh and Code

🙈 Podcast: New podcast featuring the story behind the man who married his AI. Oddly compelling to follow how this can happen, and if the protagonist is to be believed, something we will see more in coming years….

A good lesson in the potential for creators of AI tools to use psychological elements such as empathy, to both appeal and then fool users into thinking it is more than a pile of code

Did You Know? The first computer bug was literally a bug—in 1947, Grace Hopper found a moth trapped in a Harvard Mark II computer, coining the term "debugging" in the process.

Till next time,

And The Rest Is Leadership