Too much computer: Anthologie Perth’s Paul Groth on the rise of ‘AI Leadership Disease’

Paul Groth, Development Lead at Anthologie Perth, takes a pointed look at what he calls “AI Leadership Disease”, with the growing tendency for organisations to outsource thinking, writing and decision-making to artificial intelligence.

 

The government is developing a service for people suffering from AI Leadership Disease (AILD). Not to be confused with Chat Bot Syndrome (CBS), or Long-AI. The AILD pandemic has spread to 193 countries, and has infected an estimated 43 percent of the population worldwide. Promising new research from big pharma shows they could have a vaccine for AILD within 6 months.

Some common symptoms of AILD
You use AI to answer most questions, and often copy the output into group chats.

You use AI to write social posts, sometimes entire articles, and possibly using a fully automated process. You don’t read any of it yourself, but you expect other people to read it and applaud you for whatever it said.

You use AI to make a website on your computer, and you’re very proud of it. People you show it to must marvel at your genius work, while you tell them you did it… you’ve solved the problem.

Some common symptoms of the AILD-SS (Super Spreader) variant
You use AI to make a website, and manage to infect people outside of your organisation by successfully deploying it to a server where it can spread publicly.

You’re in a position of power in your organisation, and have added minimum AI usage quotas into policies and performance management criteria.

Enough of that… I think most of us are sick of hearing about AI whether it’s good, bad or otherwise. But since it’s forced upon us, we have to deal with it. And since it’s useful in some instances, you need to use it to remain competitive in the market.

AI can be a handy tool for people with the skills to do the job if there was no AI.

What makes AI a great tool, is also what makes it a disaster. That is, responding to questions and tasks that save you time, or allowing you to explore things outside of your expertise in detail. That might not make sense, but let me explain with an example.

My professional expertise is in software engineering. I often use AI to assist with diagnosing bugs, and sometimes for building something new. In any case I modify the AI output manually to fit our well-defined code standards. I run automated tests, manually check it for performance and security issues and make sure it meets requirements and compliance. It gets reviewed and tested by another engineer. Then it’s tested in other environments by non-engineers and the client. Once everything passes, it can be deployed to the production environment (live server). In production, it’s constantly monitored in real-time by various systems for security and performance, and because you can’t trust the machine for everything, it’s also monitored by real people at set intervals.

That’s a very simplified explanation of how we build robust and compliant software that people can trust and feel comfortable with at Anthologie. These processes are mostly standard practice in the industry, apart from the AI, which it’s too new and constantly changing for any industry standard practice to have taken hold. My point here is to highlight the things you don’t usually see or know about if you’re not a software engineer. With that in mind, I’m hoping you start to see how your well-intentioned actions with AI might not be as good as you think.

So what happens if you make a nice website with AI, and you’re not a software engineer, you skip most of that process, and don’t have the expertise to recognise when something might go wrong and cause harm? Nothing good.

There is a long list of potential security issues to be aware of. Vulnerabilities at the server level, vulnerabilities from malware/scams/attacks originating from anywhere else in the world simply by visiting your website, and an ever-growing number of other issues. Quick side note – did you check that code you deployed publicly for any sensitive information or API keys that might have accidentally been left in there?

Whether AI has all the context it needs or not, it can’t be trusted to generate valuable output. If you are feeding the AI everything it needs to know, that’s probably worse, plus you still can’t trust it, plus you gave it sensitive information.

Even with the development process above, building something new that has purpose, is useful, and solves a problem in a way that the majority of the target audience will accept and interact with, is not a one-person job. Before the development team gets a job, it’s been through many different skill sets, internal and external collaboration, feedback and iterations.

This post is not about software engineering… that was one example. I could have chosen any two professions, like a Neurologist setting up an AI assistant as a replacement for their receptionist… but to my point, I stuck to my area of expertise. This post is about you or someone you know, and how your usage and sharing of AI affects other people.

Example
You generated text with AI, and didn’t read it, or understand it… why would someone else bother to read or understand it? Further to that, if you generate something about a topic outside of your area of expertise and send that to a person who does have expertise in that topic… you’ve had too much computer. Go home, have a rest, and think about what you’ve done.

Example
To save time for your team, you ask AI to build the structure of a document for a project, and maybe a brief explanation of what to write in each section – to make it extra quick and easy for them. You send the document to the team to fill in the details. In most cases, this will make the job take many times longer than it would if no AI was involved at all. It will cause confusion when things don’t match previous discussions, it will have unnecessary parts, and be lacking in necessary parts. The structure will cause repetition, or break down concepts in an illogical order.

If you only have one take away from this article, I hope it’s this: Before you hit that button to share something generated by AI… think how it might affect the recipients. Maybe hit delete instead.

Use AI in moderation, and please quarantine and seek help if you think you have AILD.