> ## Content Index
> Fetch the complete content index at: https://bayesiantheory.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# AI Bayesics #1: how and where to use AI
- URL: https://bayesiantheory.com/ai-bayesics-1-how-and-where-to-use/
- Published: 2026-03-11T12:50:35.000Z
- Updated: 2026-10-06T09:02:03.000Z
- Description: If you want to learn how to use AI but don't know where the on button is, this piece is for you.
- Author: Gregg Bayes-Brown
- Tags: AI Bayesics, cognitive, AI Safety, artificial intelligence, AI risk, Training, philosophy, AI, psychology, Coaching, AI enablement, AI training, Newsletter, #Migrated-1791211413550, #Import 2026-10-05 14:44

AI has the biggest risk profile of any technology to date. Societal, economic, personal, environmental, professional, even existential. It makes nuclear weapon proliferation look like a reasonable health and safety directive.

But the main barrier - and therefore the greatest risk - is psychological. Effective use of AI requires overcoming anxiety, changing our mindset, and fundamentally challenges us to think in entirely new ways to work with the technology. This new series, AI Bayesics, addresses that challenge.

[Subscribe](#/portal/signup)

Our lack of understanding about AI breeds anxiety at all levels across an organisation. Leaders are embarrassed to put their hands up and admit they don’t know what they don’t know. Managers expected to carry to baton get minimal support and are instantly overwhelmed. Workers are worried about their jobs, the environment, and the very future of society.

You want to learn everything about AI, but it’s a bottomless pit. You turn to LinkedIn and YouTube and get bombarded with “paradigm shifts” and “inflection points” – often just others’ anxiety dressed in consultancy wonk. Every time that you learn a new tool, another one comes along and renders it redundant.

Cognitive offloading – the handing off of tasks to AI - can erode mental capability. Used improperly, AI can rob you of purpose, self-worth and challenge. This hollowing out turns to apathy and then right into depression.

This all leads to one word: burnout.

Turning that around is what this series is all about.

My goal is to turn the anxiety you feel into agency. My ambition here is to show you what’s in it for you and to provide you with the knowledge you need to harness that advantage. Through this series, you will become a self-sufficient AI native. I’m not doing the fishing for you. I’m going to show you how to fish.

We’ll be starting from the beginning so that anyone - regardless of their age, background, or current level of understanding of AI - can join in.

This will not be a course with powerpoints outdated the moment you read it. I will not be hyping the latest tools. There will be no top 10 prompts that will change your life articles. We’ll keep this free of jargon, consultancy speak, and business bullshit.

Instead, we will focus on developing the AI mindset, one which navigates uncertainty, helps you achieve greater goals, and is underpinned by psychology and philosophy. Each will cover a core component of how to work effectively with AI as a collaborator.

Let’s begin.

## It’s intelligence, stupid

Central to this series is the view that AI is not a tool to be wielded, but an intelligence to be applied.

The difference between tools and intelligence is a crucial distinction. Words have the power to transform how we perceive and interact with the world around us. It’s why I became a writer: code can hack a computer, but narrative hacks the human brain.

There are people who would object to the word “intelligence” being used to describe AI (despite it literally being called ‘artificial intelligence’). So, for clarity, this is how I define intelligence: the ability to apply knowledge to solve problems.

Note, this is not consciousness or sentience. Rather, it is a library that talks back.

The way that we effectively collaborate with AI is different from throwing around a sledgehammer. Yet, because of the ‘tool’ framing, the majority of people are hammering away. This is what Nobel Prize winner Daniel Kahneman would describe as system one thinking. Fast, automatic, unconscious decision-making with little mental effort. If someone only talks about efficiency with AI, they are trapped in system one thinking.

While there’s plenty of use cases for system one in AI, what AI Bayesics will focus on is system two. This is slower, more methodical thinking. It is thinking about thinking, known as metacognition. It prioritises divergent and lateral thinking over linear. It is taking your time rather than rushing in on gut feeling alone.

System two encourages a collaborative style of engagement with AI. This is more akin to how you engage with colleagues rather than the master-slave dynamic you have with your toaster.

We’ll be elaborating on this theme throughout this entire series. But for now, we’re going to look at it through the lens of use cases: efficiency (system one) and effectiveness (system two).

## Efficiency & Effectiveness

The first thing you need to learn about how to use AI is that it’s about reps.

You don’t go into your local gym, execute a single bicep curl, and come out looking like Arnold Schwarzenegger. Results come from putting in the work. Mastery comes from daily practise. To succeed with AI, you need to break old habits replacing them. Habits which maximise your strengths and mitigate your weaknesses.

When you go into a gym for the first time, you will be overwhelmed. What do I lift? How do I lift it? What sort of benefits am I going to get from it? Those are exactly the questions you’re about to get answers to.

What you’re going to need is any AI platform. If you’ve already got a go-to, open it up. If not, you can select any of these, each of which has a free tier. If you’re unsure and reading this at work, Copilot is likely your best bet and may already be installed on your workstation.

1. **Microsoft Copilot** — [Browser](https://copilot.microsoft.com/?ref=bayesiantheory.com) // [Windows](https://apps.microsoft.com/detail/9nht9rb2f4hd?ref=bayesiantheory.com) // [macOS](https://apps.apple.com/us/app/microsoft-copilot/id6472538445?ref=bayesiantheory.com) // [iOS](https://apps.apple.com/us/app/microsoft-copilot/id6472538445?ref=bayesiantheory.com) // [Android](https://play.google.com/store/apps/details?id=com.microsoft.copilot&ref=bayesiantheory.com)
2. **M365 Copilot** — [Browser](https://m365.cloud.microsoft/?ref=bayesiantheory.com) // [Windows](https://apps.microsoft.com/detail/9wzdncrd29v9?ref=bayesiantheory.com) // [macOS](https://www.microsoft.com/en-us/microsoft-365-copilot/download-copilot-app?ref=bayesiantheory.com) // [iOS](https://apps.apple.com/us/app/microsoft-365-copilot/id541164041?ref=bayesiantheory.com) // [Android](https://play.google.com/store/apps/details?id=com.microsoft.office.officehubrow&ref=bayesiantheory.com)
3. **Google Gemini** — [Browser](https://gemini.google.com/?ref=bayesiantheory.com) // No desktop app // [iOS](https://apps.apple.com/us/app/google-gemini/id6477489729?ref=bayesiantheory.com) // [Android](https://play.google.com/store/apps/details?id=com.google.android.apps.bard&ref=bayesiantheory.com)
4. **ChatGPT** — [Browser](https://chatgpt.com/?ref=bayesiantheory.com) // [Windows](https://apps.microsoft.com/detail/9nt1r1c2hh7j?ref=bayesiantheory.com) // [macOS](https://openai.com/chatgpt/desktop/?ref=bayesiantheory.com) // [iOS](https://apps.apple.com/us/app/chatgpt/id6448311069?ref=bayesiantheory.com) // [Android](https://play.google.com/store/apps/details?id=com.openai.chatgpt&ref=bayesiantheory.com)
5. **Claude** — [Browser](https://claude.ai/?ref=bayesiantheory.com) // [Windows](https://claude.com/download?ref=bayesiantheory.com) // [macOS](https://claude.com/download?ref=bayesiantheory.com) // [iOS](https://apps.apple.com/us/app/claude-by-anthropic/id6473753684?ref=bayesiantheory.com) // [Android](https://play.google.com/store/apps/details?id=com.anthropic.claude&ref=bayesiantheory.com)

Once you are signed in, ignore everything except the big box in the middle. This is the prompting window.

Now, write this in the prompting window:

“I am \[YOUR NAME\], \[JOB TITLE\] at \[ORGANISATION\]. I want to discover efficiency and effectiveness use cases for AI within my work. Do a background search on my profession and my organisation, and ask me any clarifying questions about what it is I do in my role. Then let’s have a conversation about how you can help me.”

On the efficiency side, you’re likely to see things that take up most of your day. In the effectiveness column, you’ll see challenges you actually want to do achieve in your role. We’re going to minimise the impact of the efficiency use cases – the stuff you don’t want to do – to maximise the strength of the effectiveness use cases – stuff you get out of bed for.

To get there, we’re going to need to free up some of your time. That’s where the reps come in.

What I want you to do is to pick 2 to 3 of the efficiency use cases. Ones with the biggest bang for your buck. Stuff that will genuinely save you some time.

Over the next week, set aside 5 to 15 minutes a day and get hands on.

Don’t worry if you don’t know how to do them. I’m not going to point you to YouTube or tell you to doomscroll LinkedIn. You’ve started with AI, and you’re going to continue with AI.

The best way to learn AI is to ask AI itself.

Simply say:

“\[Use case X\] looks interesting. Can you show me how I can set that up? Ask me any questions about how it would be most useful to me.”

Then try it out. Don’t worry - it’s okay to get things wrong. In fact, I encourage you to get things wrong. When you get something wrong, ask AI why it’s not working. Invite AI to share different ways of doing things. Getting things wrong is not failure here at AI Bayesics – it is getting data from trial and error which informs and corrects our approach.

Fundamentally, our immediate goal here to experiment. Explore a few use cases for you, get used to having back and forth conversations with AI, and working together to solve your challenges.

But for now, the most important thing is to have fun.

This Friday, we’re looking at how to communicate with AI. Despite “excellent communications skills” being a requirement of literally every job, most are failing to communicate effectively with AI and then wondering why the work’s shit.

[Subscribe](#/portal/signup)

Thanks for reading. If you found this piece useful, please feel free to share with your friends and professional contacts.

AI Bayesics will be publishing every Friday on the Bayesian Theory Substack and will be available to free subscribers. If you find this content useful, please consider supporting me with a full subscription. To sweeten the deal, we’ll be adding podcasts, videos, weekly briefings for the full tier.