Support tool or slippery slope?

Lately, I’ve noticed more conversations (and questions) about using AI for quick answers, brainstorming, and following ideas at speed, particularly from people with ADHD. I also recently saw some stats about the higher benefit of AI for neurodivergent folk beyond that for neurotypical.
On the surface, it makes a lot of sense.

AI offers instant responses, helps us unpack complex thoughts, and removes some of the friction that can make getting started feel unbearable. For many, it feels like relief. Like finally meeting their brain where it is. You have all these thoughts buzzing around your head, often not knowing how to lock some down and follow through, or where to start. But I have been pondering if that is really necessary. Is the buzzing part of processing and just noise, or should technology help us feel like we remove them from our head and action.
With this, I’ve seen a concern emerge:

Is this actually helping… or just feeding the part of my brain that wants constant stimulation?

As with most things related to technology and neurodiversity, the answer is: it depends on how it’s used.

Why AI feels so good for ADHD brains

ADHD is closely linked to differences in dopamine regulation, particularly around motivation, reward, and novelty. When feedback is slow or unclear, starting and sustaining effort can feel disproportionately hard.

This is where AI can feel like a game‑changer.

• You ask a question → you get an answer immediately
• You share a messy thought → it gets structured
• You feel stuck → something moves

That immediacy can support task initiation, reduce overwhelm, and help with idea generation, areas where executive function often struggles.

From an executive function perspective, AI can act as external scaffolding:

• Holding information in working memory
• Breaking tasks into smaller steps
• Summarising or reframing when cognitive load is high

For many people this can be the difference between not starting at all and getting just enough traction to move forward. Used this way, AI isn’t replacing thinking – it’s making thinking possible.

Where the tension starts to show

The same things that make AI useful can also become problematic.
The ADHD brain is not just sensitive to reward, it’s very sensitive to novelty and stimulation. Instant answers come with tiny dopamine hits, and it doesn’t take much for “quick help” to slide into constant checking, prompting, and idea‑jumping.

I’ve seen this show up as:

• Moving rapidly from one idea to the next without consolidation
• Using AI before sitting with uncertainty or complexity
• Feeling productive, but not retaining or integrating much

Some clinicians and researchers describe this as a learning trap: information is accessed easily, but processed shallowly. The work looks like it’s happening, yet deeper understanding doesn’t always follow. This doesn’t mean AI is “bad for ADHD.” It means unstructured AI use can amplify existing patterns, rather than compensate for them. It’s like getting a constant hit and reward, but lacking depth. I wonder how that feels for someone and impacts them long term. I could start that report and sit and contemplate, or I could ask Copilot to do the outline to help me get going. Then I might realise after the outline I can just ask copilot to write each section and its done. Hmmm… helpful or harmful?

The risk of outsourcing too much

Another concern raised in recent research around generative AI is over‑reliance.
If AI consistently:

• Starts the thinking
• Makes decisions
• Resolves discomfort

…then the brain doesn’t get the opportunity to practise those skills itself.

This matters, because ADHD support has never been about removing all difficulty. It’s about:

• Reducing unnecessary friction
• Building sustainable strategies
• Strengthening self‑trust over time

AI can absolutely support that – but only if it’s used as a tool, not a stand‑in. I cannot help but relate this to phone addiction and social media. The way we have such a habit to just pick up the phone and read the news, check messages or use apps.

We don’t realise how much it is changing our brain chemistry and our life, just like over-reliance on AI removes opportunity to use skills relevant to our daily work.

A more helpful way to think about AI use

The most balanced framing I’ve come across, and one that resonates strongly with my work in executive functions, is this:

AI works best as scaffolding, not a shortcut.

In practice, that might look like:

• Trying to think first, then asking AI to help refine
• Using AI to ask you better questions, not just give answers
• Summarising after reading, not instead of
• Using it to slow thinking down when needed, not only speed it up
   

When AI supports reflection, structure, or clarification, it often complements ADHD needs beautifully. When it replaces discomfort entirely, it can quietly undermine learning, confidence, and focus.

So… is AI good or bad for ADHD?

Neither.
It’s powerful. And power always asks for intention.

For people with ADHD, AI can:

• Reduce overwhelm
• Support initiation
• Act as an external executive function
   

It can also:

• Feed dopamine‑seeking loops
• Encourage shallow engagement
• Become another source of distraction dressed up as productivity
   

The difference isn’t the tool. It’s the relationship we build with it. As with most technology, the real work is not learning how to use it, but learning when, why, and to what purpose.

Just like with considering time without our phone, or off social media, I think we need to understand and consider our use of AI and if its helping us or reducing grey matter.