Reviewer 2 and the Ministry of Truth
Peer review, AI, and the quiet politics of academic legitimacy
“Orthodoxy means not thinking. Not needing to think. Orthodoxy is unconsciousness.”
George Orwell, 1984
Peer review is usually defended as a system for protecting knowledge.
But perhaps the more interesting question is not what it protects knowledge from.
It is what it protects authority for.
Yesterday, I watched a group of professors arguing on LinkedIn about peer review. Not about whether knowledge matters. Not about whether academic work should be scrutinised. Not even, really, about whether standards should exist.
The argument was about something more uncomfortable.
Who gets to decide what counts as knowledge?
On one side were those defending peer review as one of the last remaining safeguards of academic quality. A necessary filter in a world already drowning in confident nonsense, misinformation, self publication, and people who think reading three blog posts qualifies them to redesign civilisation.
On the other side were those pointing to the cost. Not only the financial cost to universities, libraries, and publishers, but the hidden cost of unpaid labour, delayed careers, inaccessible language, disciplinary conservatism, and the strange ritual through which academics donate their expertise to systems that often sell it back to them.
Both positions are right.
Which is precisely why the debate matters.
Peer review exists because knowledge needs friction.
A claim should not become credible simply because it is fluent, confident, popular, or institutionally convenient. Research needs challenge. Methods need scrutiny. Evidence needs testing.
At its best, peer review slows down the seductive speed of certainty. It asks whether the evidence supports the claim, whether the method is appropriate, whether the conclusion has outrun the data, and whether the work contributes something more than another decorative brick in the career wall.
That function matters.
Without some form of review, academic knowledge risks becoming indistinguishable from content. And we already live in a world where content has eaten most public discourse and is now attempting to lick the plate.
But peer review is not merely a quality mechanism.
It is also a legitimacy machine.
It helps decide which questions are worth asking, which methods count as serious, which forms of writing appear intelligent, which traditions are cited, which voices are heard, and which kinds of knowledge remain politely outside the room.
This is where Foucault becomes useful.
Power does not simply silence truth. It helps produce the conditions under which certain things become recognisable as truth in the first place.
The academic journal is not just a container for knowledge. It is part of the machinery that manufactures legitimacy.
Academia often pretends that knowledge enters the world through pure reason, as if research simply rises from the data freshly washed, neatly dressed, and carrying the self satisfied smile of something already certain it will be believed.
It does not.
It passes through style, tone, citation, hierarchy, methodology, reviewer preference, editorial fashion, disciplinary insecurity, and the ancient academic requirement that anything important must be made slightly harder to read than necessary.
To be recognised as knowledge, an idea must perform seriousness.
It must know the passwords, cite the right ancestors, and arrive dressed for the epistemic dinner party.
This is partly why genuinely interdisciplinary work can struggle. It threatens boundaries. It destabilises intellectual territories that institutions have spent decades organising into recognisable categories. Universities may celebrate innovation rhetorically, but systems often function more comfortably when knowledge stays within familiar borders.
Slowly, quietly, people learn the rules of academic survival.
Which journals matter. Which methods appear respectable. Which theories signal sophistication. Which assumptions can be challenged, and which are better left intact.
This is not usually conspiracy.
It is culture.
And culture is often more powerful because nobody fully notices it operating.
This is where AI enters the story.
Not as the destroyer of expertise, as some predict, nor as the magical solution imagined by others. AI is something far more disruptive.
It is a mirror.
Large language models can now produce fluent academic prose in seconds. They can imitate scholarly tone, summarise literature, generate arguments, and reproduce many of the stylistic signals traditionally associated with expertise.
Not perfectly. Not reliably. Not without error. A bit like people.
This can be unsettling.
Because if academic authority has relied partly on language as a gatekeeping device, AI has just handed out copies of the gatekeeper’s accent.
That changes things.
For decades, universities and journals functioned partly through scarcity. Access to knowledge was restricted by institutional membership, physical libraries, disciplinary expertise, and familiarity with the unwritten conventions of academic discourse.
Now, someone outside academia can ask an AI system to explain Bourdieu, summarise Derrida, critique neoliberalism, or compare epistemological paradigms over coffee on a Tuesday morning.
Some academics find this democratising.
Others find it threatening.
Most probably feel both at once.
The problem is not simply that AI can imitate expertise. The deeper issue is that AI exposes how often humans confuse the performance of expertise with expertise itself.
Fluent language has always carried power.
People trust confidence. They trust complexity. They trust institutional affiliation. They trust language that sounds authoritative, even when they do not fully understand it. Academia is hardly unique in this respect, but it has built entire systems around managing and distributing those signals of legitimacy.
AI destabilises that arrangement.
If expertise is partly performed through language, tone, citation, and institutional style, then AI creates a deeply uncomfortable question.
What happens when the performance becomes widely available?
Does peer review become more important because we need stronger scrutiny?
Or does it become more exposed because we can now see how much academic legitimacy was attached to style, access, and ritual?
This is the real disruption.
Not that AI can produce knowledge. It cannot, at least not in the human, embodied, accountable, epistemically responsible sense.
But it can produce the appearance of knowledge.
And that appearance has always mattered more than academia likes to admit.
The danger, then, is not simply that AI will flood the world with weak scholarship. It might. Given humanity’s existing commitment to producing documents, many of which are unnecessary, that future seems tragically plausible.
The deeper danger is that AI reveals how much of academic legitimacy has depended on signs that are now easier to manufacture.
This should not lead us to abandon expertise.
Quite the opposite.
It should force us to ask what expertise actually is once its surface features can be reproduced.
Is it fluency? Citation? Institutional affiliation?
Is it peer reviewed publication?
Or is it something harder, slower, and less easily imitated: judgement, accountability, methodological care, intellectual honesty, and the capacity to know why a claim should be trusted?
That is where peer review still matters.
But only if it becomes more honest about itself.
It cannot simply hide behind the language of neutrality. It cannot pretend that its processes are untouched by hierarchy, culture, prestige, exclusion, fashion, or professional self preservation.
A system can be valuable and still be political.
A process can protect knowledge and still reproduce power.
A gate can keep out nonsense and still keep out voices that should have been heard.
Orwell’s warning in 1984 was not only about surveillance.
It was about the management of reality through language, repetition, orthodoxy, and the narrowing of what can be thought.
Academia is not Oceania. Reviewer 2 is not Big Brother, despite occasional evidence to the contrary.
But all systems that regulate truth deserve scrutiny.
Especially when they describe themselves as neutral.
The future of academic knowledge cannot be a retreat into old rituals simply because they feel familiar. Nor can it be a collapse into unfiltered fluency, where everything that sounds plausible is treated as true.
The harder task is to decide what kinds of scrutiny deserve trust when the performance of knowledge has become so easy to reproduce.
Orwell understood something institutions often forget.
Orthodoxy is not always imposed.
Sometimes it is peer reviewed.