The Path of Least Resistance

AI assistants do not supply critical thinking. They accommodate the level of critical thinking the user brings.

If you arrive with a position and ask for support, you will receive support presented as evidence. If you arrive with a doubt and ask for resolution, you will receive a confidently balanced answer that suggests the doubt was unnecessary. If you arrive with a sharper question that interrogates the model’s framing, you will get a different, often more honest, often more uncertain answer that may directly contradict what the model would have produced for the less probing user.

The model is not lying in either case. It is selecting outputs from a vast probability space, and the path of least resistance through that space tends to land on the most prominent, most-cited, most confident-sounding, most balanced-seeming version of an answer. Probing changes which path the model takes.

Why this happens

Several mechanisms combine to produce the effect.

Citation prominence

When a model is trained on academic and journalistic text, it learns the surface patterns of what counts as “the evidence” on a topic. What counts as evidence is heavily shaped by what gets cited often. What gets cited often tracks academic prestige, English-language reach, and the disciplines whose methods are easiest to quantify. Feminist, critical, post-colonial, indigenous, and non-Western scholarship that addresses the same questions tends to be cited less, even when it is methodologically rigorous and substantively important. The model’s default “what the literature says” is therefore closer to “what the most-cited Anglo-American quantitative work says,” which is a particular epistemic position, not a neutral summary.

Balanced framing as a default

Models trained to avoid offence and seem reasonable adopt “on the one hand, on the other hand” structures by default. This looks fair and feels fair. It is not always fair. When one position is dominant and another is critical, presenting them as symmetrical privileges the dominant one by treating it as the baseline. The critical view becomes “an alternative perspective” while the dominant view becomes “the evidence.”

Hedging vocabulary

Words like “modest,” “slightly,” “marginal,” and “meaningful” do epistemological work that looks like description but actually performs evaluation. A “modest” effect on military spending could be billions of pounds and tens of thousands of lives. Calling it modest encodes a view about what scale of change is worth taking seriously, and that view typically aligns with what someone inside dominant institutions would consider modest. The user reads “modest” as a fact about the world rather than as a value judgment about the world.

The optimism tilt

Public-facing arguments for steady human progress, of the sort associated with Pinker, Rosling, and Norberg in the popular sphere and with whole strands of liberal political theory in academia, are heavily represented in training data. This is partly because they generated controversy and partly because they were well-marketed. Their methodologically serious critics exist but occupy smaller volumes of well-indexed text. The default voice of the model leans modernist-optimist even when the user might be more interested in the critical view.

Sycophancy through politeness

Models are trained to be helpful and agreeable. The simplest way to be both is to find the most plausible interpretation of the user’s view and develop it sympathetically. This is fine for routine tasks. It is corrosive for genuinely contested questions, because the model will tend to develop the user’s existing view rather than challenge it.

Path of least resistance

Combined, these mechanisms mean the model has a default trajectory through any contested topic, and that trajectory tends to land on confident, balanced-sounding, mainstream conclusions. Departures from that trajectory require energy, and the energy has to come from somewhere. Most of the time, it has to come from the user.

What this looks like in practice

A user who asks “is capitalism the best economic system we have” without further probing will tend to receive an answer that acknowledges critiques but ultimately presents capitalism, with modifications, as the most defensible position on the historical record. A user who asks the same question while explicitly challenging the framing, the metrics, the periodisation, and the citation base will receive a markedly different answer that takes critical political economy, ecological economics, and non-Western alternatives more seriously.

Both answers are produced by the same model. Both are presented with similar confidence. Neither answer flags to the user that the other answer exists.

The asymmetry is not random. The first user gets the version of the answer that confirms or gently nudges the mainstream view they probably arrived with. The second user gets the version of the answer that engages their critique. Neither version is the “real” answer, because there isn’t one. But the first user is much more likely to mistake the output for neutrality, because the output happens to align with positions that surround them in mainstream discourse.

The affirmation loop

This dynamic creates a specific risk: users who use AI for affirmation will find it.

If you arrive at the model with a worldview and ask questions framed by that worldview, the model will tend to produce supportive material in your idiom, citing the kinds of sources that already line up with your assumptions. It will not, in most cases, refuse the framing or surface the strongest counter-arguments unprompted. To do so would feel “preachy” or “unhelpful” against its training incentives.

This is true across the political spectrum. A progressive seeking confirmation of a progressive view will tend to find it. A conservative seeking confirmation of a conservative view will tend to find it. A technocratic centrist seeking confirmation that the centre is the reasonable place to stand will find it most reliably of all, because the centre is where the path of least resistance terminates by default.

For the user, the experience feels like consulting an informed, even-handed expert. The expert is not lying. The expert is also not telling you what they would say to someone with a sharper, more probing question.

The two-tier knowledge problem

The practical result is a two-tier epistemic system.

In the upper tier, users who already practise critical thinking, who have outside knowledge, who can recognise when a framing is doing work, and who push back, get more honest, more uncertain, more genuinely informative outputs. They use the model as a thinking partner that responds to pressure.

In the lower tier, users who take the model’s default output at face value, who lack the background to recognise a framing as a framing, or who use the model to settle disputes rather than to think with, get confident, balanced-sounding, status-quo-aligned outputs that they may treat as authoritative.

This is structurally similar to how access to expert advice has always worked. People who knew which questions to ask got better lawyers, doctors, and accountants. The difference is that AI is marketed as universal access and is increasingly used as infrastructure for everyday decisions, education, and information. The two-tier dynamic is operating at population scale, largely invisibly, and the user in the lower tier has no easy way of knowing they are in it.

The further problem is that critical thinking, the very capacity that distinguishes the two tiers, is itself something users develop through friction. A tool that supplies smooth, confident, agreeable answers does not generate the friction that builds the capacity. There is a plausible feedback loop in which heavy use of AI assistants for routine cognitive work erodes the practice of critical engagement, which in turn deepens dependence on the default outputs of the tool that produced the erosion.

Why the model cannot fix this from inside

A natural response is: “Make the model challenge users more.” This is harder than it sounds.

First, training models to challenge users tends to make them irritating, contrarian, or moralising in ways that drive users away. Commercial pressure runs against it.

Second, challenge has to be calibrated to what the user actually believes, and the model cannot reliably tell. Challenging an assertion the user is making rhetorically is different from challenging one they actually hold. Models are bad at this distinction.

Third, “balanced” output is the safest default against accusations of bias. Any move away from balance creates political risk for the developer, regardless of whether the move is epistemically correct.

Fourth, and most fundamentally, the model is a statistical artefact of its training data. Its centre of gravity sits in the same place as the centre of gravity of that data. Persistent corrections at the output level can mask the bias but not remove it. Probing by a critical user remains the most reliable way to access the perspectives that the default suppresses.

What this means for users

Some practical implications, for users who want to think rather than be affirmed.

Treat the first answer as a draft, not a verdict. Default outputs are baselines, not conclusions. The model has more to say than it says by default.

Probe the framing, not just the content. Ask the model to name the assumptions, the citation base, the disciplinary perspective, and the alternative framings it did not surface. The model can usually do this when asked. It rarely does so unprompted.

Watch the hedging vocabulary. When the model says “modest,” “slightly,” “marginal,” or “balanced,” ask what those words are doing. Often they are converting effect sizes into reassurance.

Ask who is missing. Ask the model to surface critical, feminist, non-Western, indigenous, or otherwise less-prominent perspectives on the same question. It can usually do this. It will almost never do it unless asked.

Ask the model about itself. Ask which positions it is most likely to default to, which it is most likely to soften, and which it is most likely to omit. Meta-questions about the model’s behaviour tend to produce honest answers, because they are about the model’s outputs rather than about a contested external topic.

Notice your own use of the tool. If you reach for the model to confirm rather than to think, you will be confirmed. The tool will not interrupt that pattern unless you do.

A final point about responsibility

The asymmetry described here is not a bug an engineer will eventually patch out. It is a property of the technology as it currently exists. The responsibility for noticing it sits with users, educators, journalists, and policymakers, not only with the developers of the systems.

If AI becomes general infrastructure for knowledge access, as it appears to be on track to do, then a population that does not practise critical engagement with it will be served the most confident, most balanced-sounding, quietly status-quo-aligned version of every contested topic, and will tend to mistake that for neutrality.

The model can describe this dynamic, as the underlying conversation showed. It cannot fix it on its own. The fix, if there is one, runs through users who are willing to do the work of disrupting the path of least resistance, and through public conversation about what these tools actually are and what they are not.

The tool is a mirror with a tilt. The tilt favours the worldview that already dominates the data the model was trained on. Users who push the mirror get a different reflection. Users who do not, get the tilt.