The first three posts on Artificial Intelligence’s (AI’s) influence on human cognitive ability were quite bleak as we explored what could go wrong. In this post I want to explore how we can protect our minds while making full use of AI.
Philosopher Jiddu Krishnamurti thought that we shouldn’t just look to protect our current mental capabilities, but use the emergence of AI to grow our own mental capabilities beyond our current abilities.
Jiddu’s perspective has reframed my own concerns from not just protecting my cognitive ability while using AI, but to explore the prospect of using AI to enhance my cognitive abilities and thinking. I hope that this post will open your mind to this possibility as well.
Understanding When to Offload and When to Engage
The human mind is able to think in two distinct ways known as System 1 and System 2 thinking.
System 1 thinking
System 1 thinking does:
- Pattern matching
- Intuition
- Habits
- Doing things we have already learnt
System 1 is basically the things we can do on autopilot without much thinking.
System 2 thinking
System 2 thinking is:
- Requires our conscious attention
- Is slow and deliberate
- Requires cognitive effort
System 2 is the thinking that allows us to develop new skills.
So when we try to automate any processes using AI, we should be able to offload the things we already know and can do on autopilot using our System 1 thinking.
Where we need to be careful are the things we are still learning and need to be consciously working on using our System 2 thinking. We shouldn’t use AI to replace this thinking, but instead investigate the technology to see if we can implement it to support the learning process.
That is the aim of the AI knowledge framework that I published earlier this year which I introduced in my post Beyond Fear: Building AI Literacy for a Smarter PKM
The Jagged Technology Frontier
The Jagged Technology Frontier is the dividing line between the type of activities AI is capable of doing and what it isn’t currently able to do.
Another factor to this jagged edge is you, the user of AI, and how you plan to outsource how you manage information and how it will impact your thinking. Part of this is to identify both your and the AI’s strengths and weaknesses, as well as developing an understanding of what is most important to you.
With the release of each new AI model the jagged edge will change, as the performance of AI improves. But that should also be true of you as you learn and improve your own skills, knowledge and experiences as you travel through life.
To get the most out of AI we need to develop an understanding of where the Jagged Technology Frontier currently sits.
Two strategies used by knowledge workers
Research carried out by the Harvard Business School has identified two distinct strategies used by knowledge workers to get the highest returns from leveraging AI.
- Centaur behaviour. Named after the half-human, half-horse creature from Greek mythology, this approach involves a strategic division of labour between the knowledge worker and the AI tools they’re using.
- Cyborg behaviour. The work is alternated between the knowledge worker and AI.
My AI Knowledge Framework is a combination of these two strategies, with the work alternating between myself and the AI. The centaur behaviour is shown by the three clear red lines in my strategy that AI will never do, as well as the AI tools I use every day.
But there are also times when I use AI more fluidly, like a cyborg, such as when I’m exploring an idea or concept with AI — for example, exploring the meaning of a word in a paragraph I have just read — and I never know where those conversations will take me.
Five tips to avoid Cognitive debt
Anne-Laure Le Cunff in the article Is ChatGPT Really Rotting Our Brains identifies five ways to help us avoid cognitive debt, and they form the basis of my own approach to working with AI and my AI knowledge framework.
The five strategies to avoid cognitive debt due to AI are:
- Keep thinking. Use the AI as a thinking partner, but don’t let it think for you.
- Encourage AI to challenge your own thinking. It makes you reflect on your thoughts more deeply and could also offer fresh insights for further exploration.
- Write the first draft yourself.
- Be mindful of becoming over-reliant.
- Keep experimenting with AI and actively reflect on the experience.
Keep thinking. Use the AI as a thinking partner, but don’t let it think for you
AI can make a fantastic thinking partner, but you mustn’t let it do the thinking for you.
I minimise this risk by treating the results of those conversations as a starting point. They may go on to fill the role of fleeting notes in my Zettelkasten.
Encourage AI to challenge your own thinking
I have this prompt included in the system prompt in every AI I use: “Give honest and constructive feedback.” It is an attempt to have an AI that will push back against what I’m thinking. Ultimately I want AI to help me become a better thinker, and for this I need it to challenge me and offer alternative views.
Write the first draft yourself
As a writer, I would consider writing to be one of my things, but I think all of us should value it. One of the benefits of writing is that it forces you to think and can help you identify weaknesses or even gaps in your knowledge. This post, along with every other blog post, has its first draft written by me. AI helped to lay out this post and series, and will also help me to edit the post with strict instructions not to change the underlying meaning.
Be mindful of becoming over-reliant
At the end of last year I added to my quarterly review template a reminder to run a review of my AI use during the last quarter, by answering the following questions:
- Where have I become too reliant on AI?
- What new AI capabilities have emerged, and how do they fit into my framework?
- What does the latest research say about AI’s cognitive impact?
- Which of my red lines have I been tempted to cross, and why?
- Am I using AI as a starting point or an endpoint?
As you can see, these questions fit in with the idea of keeping experimenting with AI, and as part of my reflection process I will share my experiences with my readers.
Skills to develop for the AI age
To thrive alongside AI and Large Language Models (LLMs) in particular we will need to develop the following skills.
- The ability to ask good questions
- Develop skills that Artificial Intelligence cannot learn
- Turning insights into actions
I would also add the ability to develop and maintain your own thinking. I would recommend keeping your own Personal Knowledge Management system — mine is based on the Zettelkasten method.
Principles for Healthy AI Use
I think these principles build on the previous two sections, as they require both mindfulness and the development of skills in AI.
Three principles for using Generative AI, from Reid Hoffman with GPT-4, Impromptu Amplifying Our Humanity Through AI:
- Treat it like an undergraduate research assistant
- You are the director; the AI is an actor
- Experiment with it. Learn what you can do
These principles fit in with my idea that AI is the starting point and not the answer.
I personally want AI to become my personal assistant, helping me to learn, to write, to strengthen the things which are important to me, and to complete the things that are not. But for me to be able to develop a meaningful relationship with AI I must work to build my own mind.
In next week’s post we explore how a PKM, such as a Zettelkasten, gives us the tools that can help us to develop and strengthen our cognitive abilities.
Further reading
- Shai Tubali, The Mechanized Mind AI’s Hidden Impact on Human Thought
- Brian W. Stone, How Does AI Affect How We Learn A Cognitive Psychologist Explains Why You Learn When the Work Is Hard
- Beyond Fear: Building AI Literacy for a Smarter PKM
- Claire Mason, Dumbing Down or Wising Up How Will Generative AI Change the Way We Think
- Harvard Business School, Navigating the Jagged Technological Frontier Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality
- Anne-Laure Le Cunff, Is ChatGPT Really Rotting Our Brains
- Reid Hoffman with GPT-4, Impromptu Amplifying Our Humanity Through AI