The more choices you give someone, the longer it takes them to decide — and the more likely they are to give up entirely.
Think about the last time you typed something into Google before AI Overviews existed. You'd land on a results page with ten blue links. You'd scan titles, try to guess which would answer your question, click one, realise it wasn't quite right, hit back, try another. Thirty seconds for a question with a one-sentence answer.
That friction is exactly what Hick's Law predicts. Decision time grows with the number of choices — and ten roughly-equal options is one of the most expensive decision scenarios a user can face. The arrival of AI summaries in search isn't just a feature change. It's a fundamental reduction in choice load applied at scale, to billions of queries a day.
In 1952, psychologists Hick and Hyman gave participants a row of lights, each paired with a button. They varied how many pairs there were and measured response times. Results were clean: decision time grew logarithmically as choices increased. The first few extra options had by far the biggest impact.
Cutting from ten options to five has more impact than cutting from fifty to forty-five. This is why removing a single nav tab, pre-selecting a payment method, or collapsing search results into one answer all have disproportionate effects on how fast users move.
“Ten blue links was never a good design. It was just the only option we had. AI finally fixed it.”
For decades, Google's answer to every query was the same: here are ten links, you decide. The AI Overview changes that completely. Instead of ten roughly-equal options demanding evaluation, you get one synthesised answer at the top. Decision time drops to near zero.
Same query, same information — but one puts the work on you, and one doesn't.
Six results, all plausible. You have to click each to find the answer.
Answer synthesised at the top. Sources available if you need depth.
The second screen hasn't removed the links — it's moved them below a direct answer. Users who need the source can still find it. Users who just want the answer don't have to choose between six roughly-equal options to get it.
It's not just search. The same problem existed in every powerful tool with a complex interface — advanced query builders, filter systems, settings panels with dozens of toggles. Natural language AI interfaces collapse that load to one: a blank input and a cursor.
Full control — but enormous cognitive cost before seeing a single result.
Type what you need. The system figures out the filters.
The query builder isn't broken — for power users who need precise control, it's the right tool. But for everyone else, the AI interface gets to the answer faster because it's removed every intermediate decision.
Even without AI, Hick's Law determines whether someone completes a checkout. Payment screens are one of the clearest test cases. A user who has already decided to buy something just needs to pay. Every extra choice is pure friction.
Which is fastest? Does Klarna affect credit? What fees does PayPal charge?
Visa, PayPal, Klarna and 4 more under
“Pay another way”
Fastest method pre-selected. Other options are one tap away if needed.
The second checkout doesn't remove payment options — it pre-selects the fastest and most secure one available on the device and moves the rest behind a single overflow.
When the interface collapses to a single input, the surface area for Hick's Law goes to zero. No toggles to choose between, no filters to configure — just a question and a synthesised answer with sources.
A single prompt replaces filters, tabs, and source evaluation.
Hick's Law shows up everywhere — navigation bars, onboarding forms, pricing pages, filter systems. The most important shift happening in product design right now is AI reducing option counts at the system level — converting ten results into one answer, a twelve-field form into a single prompt.
Hick, W. E. (1952). On the rate of gain of information. Quarterly Journal of Experimental Psychology. · Hyman, R. (1953). Stimulus information as a determinant of reaction time. Journal of Experimental Psychology.