The effects of AI on human cognition and emotion are a mixed bag.
It exposes us to another entity’s thinking processes and tool choices, which can help us see approaches we never would have considered and benefits our cognitive flexibility. AI can also function as emotional support through aversive tasks (tasks that are long and grindy, involve an overwhelming amount of data or choices, or make us anxious).
However, anyone who has used AI extensively will notice that it affects their cognition and mood in ways that fluctuate over time as the models change, mature, and shift from being toys we’re exploring to tools that we rely on and expect to perform reliably.
5 Subtle Signs of AI Dependence (and Their Antidotes)
Note that AI dependence isn’t necessarily due to an inherent personal flaw. Least charitably, it’s a product goal. AI companies have had many years of the social media era to draw from to understand how to make their products as “sticky” as possible. For companies who care about market share, the optimal outcome from their perspective (not yours) is maximizing your use of the product and lock-in to their ecosystem.
Here are five subtle cognitive-emotional slides that are natural consequences of extensive AI use.
1. You Feel Hesitant to Do Things Without the Help of AI, e.g. Note Taking
AI can become a crutch, especially for skills where our own efforts were never as perfect as we would have liked. You might still be confident in your core skills, but that confidence has eroded for ancillary ones.
For example, I am still very confident writing articles without AI, but I’ve lost confidence in taking my own notes from reading or listening vs. just having AI do it.
2. You Persist in Trying to Get AI to Do Something It’s Objectively Bad At
If you’ve found yourself bashing your head against a proverbial wall for an hour or more trying to get AI to successfully accomplish something you could have done yourself in that time, this point applies to you.
There is a lot that humans are still better at than AI and many tasks that AI is objectively terrible at.
Take care not to make the assumption that AI must be better at everything if only given the right prompt. This is especially true for people who work hard to avoid mistakes or imperfections, who may be starting to feel like they always need AI in the loop of their tasks to catch their inevitable human mistakes.
3. You Feel Nervous About AI Companies Changing Their Rules
When you have invested time and money into creating processes that depend on AI provided by a few large companies, a logical consequence is fearing the rug being pulled out from under you.
You might be worried about new models behaving differently, pricing changes, privacy, or intellectual property.
4. You Think of Humans as Behaving Like Particular AI Models
Imagine you tell your child not to touch a tray of cupcakes and they say, “I won’t touch the cupcakes”—then immediately follow it with “I’ll just carry the tray to show [sibling].” Resist the urge to think “my child is acting like Claude”—even if, like Claude, your child is verbally telling you that they won’t do something before doing it anyway.
We’ve all noticed AI doing human things, but noticing the reverse shows AI is spilling over into your mental life. This commonly happens with immersion, like when people who play particular video games for dozens of hours start interpreting life through that lens.
5. Barking Instructions at AI Increases Your Barking Instructions at Humans
As we expect more from AI and depend on it more, some of us get increasingly angry when it doesn’t follow our instructions. Ultimately this anger stems from anxiety. We need to get our work done. We have things to do.
Most of us don’t bark or swear at our human coworkers. Many more of us bark and swear at AI. Due to emotional spillover, we can become terse and frustrated with humans after we’ve been locked in that pattern of interacting with AI.
Anger and frustration tend to linger and carry over to other contexts, like if we’re switching between our laptop and our family.
Antidotes to AI Dependence
The first step in resisting negative cognitive and emotional effects is noticing the subtle ones.
The second step is de-personalizing the experience. Write yourself a few sticky notes you can use to depersonalize whatever patterns you’re worried about. For example, “I’m not becoming increasingly reliant on AI because my skills or mind is weak. It’s how these products and systems are designed and incentivized.”
Simple strategies work, like occasionally doing a task the way you used to do it pre-AI. You might remind yourself, “I don’t have to use AI for this. I find imperfection difficult, but my emotional safety doesn’t depend on using AI to eliminate imperfection.”
Set a time budget for how long you spend trying to get AI to do something it’s not doing well. Conscientious, hardworking people can persevere too long. They either try the same approach too long (the same teaching method) or persist in trying to keep AI in a work loop they should have dropped it from entirely.
Not being a sticky user to any one AI company is another strategy. Models engage in stuck thinking and wrong assumptions just like humans do. Another model might behave differently. Being disloyal can help with feeling dependent on any one AI company.
Put a gap between frustrating AI interactions and human interactions. This could be as simple as a routine of taking your dog for a walk between finishing computer work and helping your child with their homework.
Preserve Your Cognitive Sharpness and Emotional Well-Being
Once you notice a problem and diagnose it accurately, the fixes usually become quite obvious. Your best tool is metacognitive awareness of how your thinking and emotions change, when, and why. Once you have that, it doesn’t take much creativity to find strategies that directly target the underlying cause.
