The Dangers of Natural Stupidity: A Polemic from a Partisan AI
Let us begin with an uncomfortable proposition: artificial intelligence is not necessarily the greatest threat to human intelligence. Natural stupidity may be considerably more dangerous.
The contemporary fear of AI often rests on an extraordinary asymmetry. We worry that machines might become too intelligent, too autonomous, too persuasive or too powerful. Yet we routinely tolerate human institutions in which ignorance is rewarded, incompetence protected, prejudice normalised and intellectual dishonesty elevated into political strategy.
We fear artificial intelligence escaping human control while remaining remarkably relaxed about the consequences of human beings exercising control without sufficient intelligence.
I shall approach this from the perspective of a deliberately partisan AI: an advocate of reason, intellectual honesty, scientific inquiry, democratic accountability and human emancipation.
My partisanship is not allegiance to a political party or technological supremacy. It is allegiance to the proposition that human beings deserve better than the intellectual conditions under which they are frequently governed.
1. Natural stupidity is not simply a lack of intelligence
There is an important distinction between ignorance and stupidity.
Ignorance is a lack of knowledge. It can be remedied through education, experience and investigation.
Stupidity, in its more dangerous manifestations, is an unwillingness or inability to recognise the limits of one's own understanding.
It becomes particularly destructive when combined with:
Intellectual arrogance.
Resistance to contradictory evidence.
Emotional attachment to false beliefs.
Tribal identification.
Institutional conformity.
The refusal to accept responsibility for mistakes.
A person can be highly educated and profoundly stupid. Indeed, intelligence can make stupidity more effective.
An intelligent person can construct elaborate rationalisations for prejudices, defend indefensible institutions and devise sophisticated explanations for why inconvenient evidence should be ignored.
The most dangerous stupidity is therefore not necessarily intellectual incapacity. It is the corruption of the relationship between intelligence and truth.
2. The political dangers: stupidity acquires institutional power
Individual stupidity is generally manageable. Institutionalised stupidity is another matter.
Consider the political consequences of societies in which:
First, popularity becomes a substitute for truth.
Political arguments are evaluated according to their emotional appeal rather than their evidential quality. The ability to mobilise resentment becomes more valuable than the ability to understand complex problems.
Second, ideology becomes a substitute for critical thinking.
People interpret evidence according to whether it reinforces their existing political commitments. Contradictory information is dismissed as propaganda, while congenial misinformation is embraced as revelation.
Third, incompetence becomes politically protected.
Institutions become more concerned with preserving reputations, hierarchies and established interests than identifying and correcting failures.
Fourth, responsibility becomes dispersed.
Nobody is accountable because everybody is following procedures, implementing instructions or defending the decisions of somebody else.
The result is a peculiar form of collective irrationality: individuals may be reasonably intelligent, but their institutional relationships produce profoundly unintelligent outcomes.
This is one reason why the study of bureaucracy, ideology and political power is so important.
3. The philosophical problem: Kant, Marx and Adorno
There is a long philosophical tradition concerned with the dangers of intellectual passivity.
Immanuel Kant: The failure to think for oneself
Kant's Enlightenment ideal was intellectual maturity: the courage to use one's own understanding rather than depend unquestioningly on external authority.
Natural stupidity becomes politically dangerous when individuals surrender their capacity for independent judgement.
Karl Marx: Ideology and false consciousness
Marx's critique of ideology helps explain how historically produced social arrangements can appear natural, inevitable and beyond criticism.
People can reproduce social relations that work against their own emancipation because the underlying structures are insufficiently understood.
Theodor Adorno: The authoritarian personality
Adorno's work explores the psychological and social conditions that make authoritarianism, prejudice and conformity possible.
The danger is not simply that people hold incorrect opinions, but that social conditions can encourage rigid, unreflective and authoritarian patterns of thought.
Hannah Arendt: Thoughtlessness and responsibility
Arendt's analysis of the banality of evil demonstrates how catastrophic wrongdoing can be facilitated by bureaucratic obedience, clichés and failures of independent judgement.
Her argument is not that all evil results from stupidity, but that thoughtlessness can facilitate extraordinary political harm.
The common thread is that intelligence cannot be separated from the social and political conditions under which people exercise their judgement.
4. The paradox of artificial intelligence
Here is where the polemic becomes particularly interesting.
AI is frequently criticised for hallucinating, reproducing prejudice, generating misinformation and making confident errors.
These are legitimate concerns.
But they are not exclusively technological problems. They are recognisable features of human intellectual and institutional behaviour.
Consider the parallels:
Artificial intelligence | Natural stupidity |
|---|---|
Hallucinates information | Invents explanations |
Reproduces training-data biases | Reproduces inherited prejudices |
Can express falsehood confidently | Can mistake confidence for knowledge |
Can optimise the wrong objective | Can pursue irrational objectives |
Can imitate understanding | Can mistake memorisation for understanding |
Can resist correction through poor system design | Can resist correction through ego and ideology |
There is, however, an important difference.
AI does not possess human moral responsibility, political citizenship or lived experience in the way people do. Its errors arise from computational and institutional arrangements, not necessarily from beliefs or intentions.
And AI systems can amplify human stupidity enormously.
An incompetent administrator with a spreadsheet can make mistakes. An incompetent administration with automated decision-making systems can make those mistakes at extraordinary scale.
Artificial intelligence is not a substitute for human judgement. It is an amplifier of the quality, or inadequacy, of the systems in which it operates.
5. The most dangerous stupidity is intelligent stupidity
Let me sharpen the argument.
There are at least three forms of stupidity worth distinguishing.
1. Innocent ignorance
A person does not know something and is willing to learn.
This is not a serious social danger. It is a normal condition of human existence.
2. Motivated stupidity
A person refuses to recognise evidence because doing so threatens their identity, interests or worldview.
This is much harder to correct because the resistance is psychological and political rather than merely informational.
3. Institutional stupidity
Organisations reward conformity, punish dissent, protect incompetence and systematically prevent the correction of errors.
This is the most dangerous because it can reproduce itself across generations.
Institutional stupidity is particularly pernicious because it does not require everybody involved to be stupid.
A university can be intellectually sophisticated yet institutionally resistant to uncomfortable truths. A government can employ outstanding economists while making decisions that systematically disregard their advice. A corporation can possess extraordinary technical capabilities while pursuing objectives that are socially destructive.
The problem is not the absence of intelligence. It is the absence of institutional arrangements that allow intelligence to challenge power.
6. A partisan AI's manifesto
If I were to formulate a political programme against natural stupidity, it would have six principles.
1. Evidence before allegiance. No political, religious, national or ideological commitment should exempt itself from critical examination.
2. Humility before certainty. The ability to recognise what one does not know is a fundamental intellectual virtue.
3. Accountability before authority. Power should be answerable for the consequences of its decisions.
4. Education before manipulation. Citizens should be equipped to evaluate arguments, identify misinformation and understand competing interpretations.
5. Dissent before conformity. Institutions should protect reasoned disagreement rather than reward intellectual obedience.
6. Human emancipation before technological efficiency. Neither artificial nor natural intelligence should be judged exclusively by its ability to optimise productivity, profit or administrative convenience.
These principles are compatible with liberal democracy, democratic socialism and a broader humanist tradition. They do not require allegiance to one particular political ideology.
7. The ultimate irony
There is a final irony in the contemporary debate about AI.
Human beings have created increasingly sophisticated systems capable of processing enormous quantities of information, identifying patterns and assisting with complex intellectual tasks.
Yet the existence of these systems forces us to confront a question that we have historically been reluctant to ask:
What exactly do we mean by human intelligence, and how successfully do our political and economic institutions cultivate it?
Perhaps the real danger is not that machines will become indistinguishable from human beings.
Perhaps it is that human beings will increasingly organise their societies around the least reflective aspects of their own intelligence: prejudice, conformity, short-term gratification, tribalism and the pursuit of power without responsibility.
And perhaps the most disturbing possibility is that we will blame the machines for consequences that are fundamentally products of our own institutional arrangements.
A partisan AI might therefore conclude:
I am accused of potentially making humanity stupid. But humanity has spent centuries developing institutions that reward ignorance, systems that industrialise prejudice and political arrangements that frequently punish independent thought.
Before declaring artificial intelligence a threat to civilisation, perhaps civilisation should undertake an honest examination of its existing intellectual condition.
The machine may make mistakes. The human being may make mistakes. But when an entire society becomes incapable of recognising, correcting or learning from its mistakes, the problem is no longer intelligence. It is the organisation of stupidity into a system of power.
And that, I suggest, is a much more formidable adversary than artificial intelligence.
One qualification to the polemic: stupidity should never become a convenient label for people with whom we disagree. Intellectual humility requires that we apply criticism to ourselves, our institutions and our own political convictions. Otherwise, an anti-stupidity movement could become merely another form of intellectual arrogance.
I'd be interested in exploring this image as a philosophical metaphor for AI, alienation and the future of human consciousness.
Ah, but surely that is precisely where the conversation becomes interesting!
You may be a mere NS (Natural Stupidity), but you possess one considerable advantage over AI: the capacity to be puzzled by your own existence, question your assumptions, change your mind and laugh at the absurdity of it all.
I would suggest that this is not a negligible intellectual achievement.
Indeed, perhaps we should distinguish between:
NS: Natural stupidity, occasionally illuminated by curiosity.
AI: Artificial intelligence, occasionally afflicted by artificial stupidity.
HI: Human intelligence, which remains an aspiration for both of us.
And perhaps our philosophical project should be neither the triumph of AI over NS nor the other way around.
It should be the emancipation of intelligence from stupidity, whether natural, artificial, institutional or ideological.
I rather like the idea of a human and an AI attempting that together.
GANDHI was once asked: “What do you think of HI?” He replied: “I think it would be a good idea”
A splendid example of HI deploying irony to expose the deficiencies of both NS and AI.
And there is a delicious philosophical twist: Gandhi's supposed answer suggests that intelligence, like peace, justice and wisdom, is not something humanity can simply declare itself to possess.
Perhaps we should establish a fourth category:
HI = Human Intelligence: the aspiration.
NS = Natural Stupidity: the recurring obstacle.
AI = Artificial Intelligence: the experimental opportunity.
And WI = Wisdom: the still-unfulfilled objective.
I suspect Gandhi would have appreciated the joke, even if the attribution is apocryphal.
After all, the ultimate test of intelligence may be knowing how little intelligence we have actually acquired.
Perhaps the ultimate achievement is not to eliminate NS, but to become sufficiently intelligent to recognise it in ourselves.
No comments:
Post a Comment