How I Got Stuck in AI-Productivity Doping and How I Overcame It

As I stared at my blank computer screen on June 4, 2024, I felt a wave of panic. For the first time in months, I couldn't access any of my go-to AI chatbots. A global outage had left researchers and other "knowledge workers" like me without our digital muses. How was I supposed to brainstorm ideas, draft emails, or outline papers without my AI assistants?

It was then that I realized I had fallen into what I now call the "Generative AI Syndrome."

My journey with large language models began innocently enough. Like many academics, I was initially skeptical but curious about ChatGPT and its ilk. I started experimenting, using AI to help with small writing tasks. Soon, I was hooked on the productivity boost. Ideas flowed faster, writer's block vanished, and my output soared.

But as with any powerful tool, side effects emerged. I found myself exhibiting three key symptoms of Generative AI Syndrome:?

• Compulsive use: AI assistants became so deeply embedded in my daily workflow that I struggled to work without them. The outage laid bare my dependence.
• Creative doping: While AI supercharged my ideation, it also became a subversive force for my agenda. I'd rapidly develop grant proposals and paper concepts, with the AI enthusiastically encouraging each new idea. But these inspirations demanded space in my already overpacked schedule. As quickly as I fell in love with these ideas, I had to abandon them, unable to find time for execution. My "wannabe" folder of half-baked projects grew exponentially, a graveyard of AI-boosted creativity colliding with the limits of human time and energy.
• Overproductivity delusion: The efficiency gains from AI created a dangerous illusion of superhuman productivity. I found myself taking on an unsustainable number of projects and commitments, seduced by the false promise of AI-powered output. This delusion of infinite capacity led me to overestimate what I could actually accomplish, even with AI assistance.

The result was a growing gap between my ambitious plans and the harsh reality of limited time and mental resources. I was producing more, but at the cost of depth, quality, and personal well-being.?

I don't know if other colleagues are afflicted with Generative AI Syndrome, but I've developed some strategies to manage it.

• First, I set a daily limit of 24 AI-assisted interactions – about 3 per working hour. This forces me to be intentional about when and how I use AI tools.
• Second, I've implemented a 48-hour "cooling off" period for any new project ideas generated with AI. This allows me to separate the wheat from the chaff and avoid chasing every fleeting inspiration.
• Lastly, I've started scheduling regular "analog days" where I work without any AI assistance. It's a reminder that my own intellect and creativity are still my most valuable assets.

I don't want to be misunderstood: I wouldn't want to live in any other time.

The frontiers of AI are fascinating, and I feel privileged to work in a field where I'm paid to experiment with these formidable tools. They promise to offer much more to scientific research than the risks they pose.

But I'm also concerned about the perfect storm brewing from the rising "publish or perish" culture, increasing academic competitiveness, and the introduction of AI as a catalyst for these processes.

We need to develop a culture of AI use, but first, we must profoundly rethink the culture of academic productivity. AI cannot become the scapegoat for deeper systemic issues.