
Summary
A fast AI answer can shape more than confidence in advice. The research described explores how expectations about effort and speed influence perceived accuracy, temporal distance and choices between immediate and delayed rewards.
At a glance
AI response speed can influence perceived accuracy and how distant the future feels. The research described links AI advisers with greater preference for smaller immediate rewards under some conditions. These effects depend partly on perceived task difficulty and whether response time signals processing speed or recommendation quality; they do not establish objective accuracy or literal control over thinking.
Is AI Gradually Taking Control of Your Thinking?
The library or the café? You cannot decide, so you turn to ChatGPT and DeepSeek for help. You send off the question, then become so caught up in their apparent deliberations that you lose yourself in the decision all over again. But how much should you trust the advice?
Your AI Is Responding…
Suppose a friend answers your question the instant you finish asking it. What would you make of that answer? Much depends on the question. In one study, researchers asked participants whether they would hand in a wallet they had found full of cash. People who took longer to decide were perceived as less honest. In that context, a pause suggested hesitation: perhaps the person was considering keeping the money. The delay therefore raised doubts about their honesty.
For tasks that require effort, however, a quick answer can seem less sincere. Take the calculation 22 × 71 × 91 + 33 × 65. While you are still struggling with the arithmetic, someone with no training in speed calculation gives an answer a second after hearing the problem. Your first thought might be, “That must be a guess.” You suspect they have picked a number rather than worked it out carefully. Here, taking longer suggests greater effort, which in turn makes the answer seem better.
What happens when AI is doing the answering? In research involving prediction questions, slower human responses were judged more accurate. For AI, the pattern ran in the opposite direction: slower responses were judged less accurate. Why should the same delay carry such different meanings?
To investigate, the researchers introduced three questions during the prediction task: 1. Is predicting the future an easy or difficult task for AI, or for a human? 2. How much effort did the AI or human put into making this decision, on a scale from 1, very little effort, to 7, a great deal of effort? 3. Considering how quickly the AI or human responded, how accurate do you think the prediction is?
Participants interpreted slower human responses as evidence of greater effort, and associated that effort with a more accurate prediction. They also saw slower AI responses as involving more effort. For AI, however, they did not connect that additional effort with greater accuracy.
Now return to the library and the café. If you asked me for advice and I spent a while weighing the choice, you might take that as a sign that I cared and could be trusted. If DeepSeek kept loading for three minutes before producing an answer, you might simply throw up your hands in frustration. The wait invites a different interpretation depending on who—or what—is answering.
Task difficulty is an important part of how we judge effort. Some tasks seem easy for AI but difficult for people; others present the reverse picture. Consider the contrast between AI struggling with image recognition and people struggling with complex calculations that combine several kinds of information. Could perceived difficulty also change how response speed shapes judgments of accuracy?
Before participants assessed prediction accuracy, the researchers told them that either a great deal or very little supporting information was available, manipulating how difficult they expected the prediction to be. A similar pattern emerged for human and AI respondents: on difficult tasks, longer response times were associated with higher perceived accuracy; on easy tasks, the relationship reversed.
So far, the discussion has concerned objectively measurable response times. But what if AI also changes how time itself feels to us?
Time May Feel Different to You and Me
Researchers have found that associating AI with fast, efficient responses can influence subjective time perception: our personal experience of objectively measurable time. At a party where you know nobody, or while waiting for your grades to be released, you may feel so uncomfortable that every minute drags. Yet a long-awaited holiday can seem to vanish in an instant.
These differences in how time feels can affect behavior. The longer a wait seems, the more impatient people can become and the more they may prefer an immediate reward. Even the units used to describe a delay can matter: “2 weeks” has been found to feel farther away than “14 days,” despite representing the same duration, and to be associated with greater impatience. Similarly, the farther away retirement feels, the less willing people are to save for it. As impatience grows, receiving a reward sooner becomes more appealing.
To examine this in an AI setting, researchers invited participants to chat with either an AI or a human adviser about holiday choices. A credit-card cashback offer gave those booking their trip that day a choice: receive $30 immediately or $35 after four weeks. While making that choice, participants also rated how distant four weeks from today felt on a scale from 1, near, to 7, far. The researchers accounted for additional variables, including participants’ attitudes.
The researchers found that interacting with AI activated a belief in rapid processing. As AI answered quickly, people felt that time was flowing faster, which made the future seem farther from the present. That encouraged them to choose a reward they could receive sooner, even when it was smaller than the delayed reward. This faster flow was a subjective experience; objective time had not accelerated.
The resulting wait can feel like this: “Is it time yet?” “Not yet.” “Is it time yet?” “Not yet.” “This is taking forever!” The expected moment still has not arrived, and the future feels farther away.
To investigate the causal relationship between time perception and preference for earlier rewards further, the researchers changed how participants understood response time as a signal of decision accuracy. The earlier findings pointed to an expectation that faster AI responses are more reliable. In the follow-up experiment, however, participants were told that AI took longer because it was comparing different strategies to provide better advice. This explanation challenged the assumption that a quicker AI answer necessarily meant a more dependable one.
When response time was understood as an indicator of processing speed, AI advisers increased participants’ choices of immediate rewards compared with human advisers. When participants understood that time reflected the quality of the travel recommendation, this effect weakened. Giving people another basis for judging answer quality could therefore change the meaning they attached to AI’s speed.
Living With AI
These findings suggest that we think of algorithms as more than tools. We associate them with particular qualities, including speed and objective analysis. Those associations can shape how credible information seems and become part of the context influencing our opinions and choices. AI’s rapid answers can also affect how we judge time and how distant the future feels.
That raises several questions.
- If a conversation with AI about buying an expensive item changes our sense of time and makes the future feel farther away, might a short-term loan seem preferable to a long-term one?
- Might people complete assignments from a human teacher on time, yet feel they have more time for assignments from an AI teacher and therefore procrastinate more?
- And if the future feels more distant, might people pay more attention to the present?
Researchers have identified many such questions for future investigation; they have yet to establish the answers. As we watch AI develop, we too must learn how to live alongside it. Beyond assessing what AI says and how quickly it responds, we need to retain responsibility for our own choices and thinking. How we balance AI, time and the future is a question we will have to face.
Frequently asked questions
Does a faster AI response mean the answer is more accurate?
The findings described concern perceived accuracy rather than measured prediction accuracy. Perceived task difficulty changed the pattern: longer responses appeared more accurate on difficult tasks, while the relationship reversed on easy tasks.
Can interacting with AI change how time feels?
In the travel-adviser research described, AI interaction activated beliefs about rapid processing, making time seem to flow faster and a future reward feel farther away. This was a subjective experience; objective time did not accelerate.
What weakened the preference for immediate rewards with AI advisers?
The effect weakened when participants understood response time as reflecting travel recommendation quality. When response time signaled processing speed, AI advisers increased immediate-reward choices compared with human advisers.
Does the article establish that AI causes procrastination or changes loan preferences?
No. Changes in loan preferences, assignment procrastination and attention to the present are presented as questions for future investigation, with answers yet to be established.


