Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Clear out junk files and repair common Windows errors3Scan for outdated or missing drivers - takes under a minuteIf AI makes judgments cheaper and easier to obtain, people may ask for more of them—but that possibility is not yet an established Jevons paradox. The evidence behind the original paradox concerns energy and resource use. Research on AI, decision-making, and cognitive effort offers relevant clues, but it does not show that AI has caused total human judgment to rise, or that it has weakened unaided judgment over time.
What is the Jevons paradox?
William Stanley Jevons discussed the idea in his 1865 book The Coal Question. More efficient steam engines used coal more effectively, but that improvement could also make coal-powered applications more economical and widespread. The result Jevons focused on was not simply greater efficiency per engine: it was the possibility that total coal consumption could increase. Blake Alcott’s historical account explains the concept in “Jevons’ paradox”.
As an Amazon Associate I earn from qualifying purchases.
Modern analysis distinguishes rebound from backfire. Rebound occurs when lower effective costs lead to extra use that offsets some expected savings. Backfire is the stronger case: use rises enough to exceed the expected savings, so total consumption is higher than in the relevant counterfactual. Neither outcome follows automatically from an efficiency improvement. Steve Sorrell’s 2009 review says that testing for backfire is difficult and that the evidence was far from conclusive, while suggesting economy-wide rebound may be larger than commonly assumed. Read the review in Energy Policy.
The United Nations Development Programme’s Human Development Report 2025 says: “Evidence from dozens of studies suggests that economywide rebound effects following energy efficiency gains exceed 50 percent, on average.” That estimate concerns energy-efficiency studies. It is not a measure of judgment, AI decision quality, or cognitive skill. Read the report.
#1 Best Overall
If AI makes judgment cheap, will we use more of it?
That is a reasonable hypothesis, not a demonstrated law. An AI tool might reduce the time or effort needed to generate a recommendation, assessment, or ranking. A lower cost could make people use such outputs in more situations, including ones where they previously relied on their own judgment or did not make a formal judgment at all. But to call this a Jevons-style effect, evidence would need to show that the increased volume or resource use offsets an expected saving under a clear counterfactual.
It matters what “more” means. More AI queries, more decisions made, greater compute use, and more consequential choices are different outcomes. Additional judgments might meet a real need, or they might simply increase low-value decision volume. A rise in AI use alone does not establish backfire in human judgment.
Rank #2
What does the evidence say about AI and judgment?
AI environmental rebound is an active question
Efficiency improvements can reduce the cost or energy required per computation, while greater demand or more complex models could offset some of those marginal savings. The UNDP report discusses this possibility. A 2025 FAccT paper by Alexandra Sasha Luccioni, Emma Strubell, and Kate Crawford groups AI rebound effects into material or physical, economic, and social or behavioral effects. It also notes that direct comparisons and impacts remain under-explored. These works establish an important research concern, not a settled finding that AI has produced net rebound in every setting. Read the FAccT paper.
Knowing choices train AI can change behavior
A 2025 Proceedings of the National Academy of Sciences paper reports five experiments using an ultimatum-game task. Participants who knew their choices would train AI became more punitive toward low offers than control participants. The change persisted in a later task that was no longer used for training. The authors write, “However, our work challenges this assumption.” The finding shows that an AI-training context can affect behavior in those experiments; it does not show that people ask AI for more judgments, that total decision demand rises, or that unaided judgment degrades over time. Read the study.
Cognitive offloading is a possible mechanism, not proof of skill loss
Research on cognitive demand describes how people use simplifying strategies and shift some control demands to their environment. This offers one way to think about how tools might change mental effort. The 2010 paper, however, is not a study of current generative AI and does not establish long-term loss of judgment skill from using it. Read “Decision Making and the Avoidance of Cognitive Demand.”
Taken together, these sources do not establish a judgment-specific Jevons effect. They offer adjacent evidence and plausible mechanisms, but not a direct longitudinal or demand-elasticity finding that AI increases total human judgment enough to constitute backfire, or that it erodes unaided judgment over time.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Does AI make people think less for themselves?
It can change where effort is spent, but the available evidence here does not settle whether people think less for themselves overall. Delegating a routine assessment might free attention for other work; relying on an output without examining it might reduce personal scrutiny in a particular decision. Neither example, by itself, demonstrates a lasting change in skill or a general decline in independent thought.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
To test that stronger claim, a study would need to distinguish immediate assistance from later unaided performance and account for what people would have done without the tool. A change in behavior while a person knows their choices are training AI, such as the one observed in the PNAS experiments, is not the same as a lasting loss of judgment ability.
Best Value
How to assess a claim that AI has caused judgment rebound
Before accepting a claim that easier judgment has led to more judgment—or worse judgment—check what was measured and against what alternative. A useful comparison should specify:
- Outcome: total decisions, AI queries, compute or energy use, decision quality, time saved, or unaided skill retention. These measures cannot stand in for one another.
- Counterfactual: what people would have decided, how much time they would have spent, or what resources they would have used without the efficiency gain.
- Time horizon: an immediate shift in tool use may differ from longer-term changes in demand or habit.
- Cost bearer: whose time, money, compute, or exposure to risk changes—and whether a saving for one party creates a cost for another.
- Value of added use: whether extra judgments address previously unmet needs or mostly add low-value decisions.
- Evidence design: whether the result is a causal finding, a correlation, an illustrative example, or a theoretical possibility.
These distinctions also help keep three questions separate: whether AI changes resource consumption, whether it changes how many decisions people make, and whether it changes decision quality or unaided skill. Evidence for one does not answer the others.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Recommended Free Tools
