The idea that artificial intelligence (AI) is facing organised or diffuse backlash—from creatives, labour unions, civil-society critics and even existential-risk advocates—is more than a niche grievance. In this article I look into why people are hating on AI, assess whether those concerns hold water, examine the specific fear of a future general or super-intelligent AI (AGI/ASI) challenging humanity’s survival, and consider how AI (and its creators) might mount a reasoned defence. Our view: the backlash is partially justified, and how the AI world responds will play a defining role in its legitimacy and future.
1. Why is there a backlash? The roots are varied. Many feel the arrival of generative AI (art, writing, music) threatens human uniqueness: a recent study found that people’s dislike of AI-generated art was linked to the feeling that “digital works threaten the last fortress of human supremacy.” [2] Moreover, generative AI tools have proliferated while many of their harms fall disproportionately on marginalised communities—labour shifts, environmental burdens, biased data. [6] Public polls likewise show large-scale concerns: one survey found 72% of U.S. adults said they had major doubts about AI, citing privacy, surveillance, algorithmic bias and lack of transparency. [17] In short: many see AI as a power shift—jobs displaced, control lost, creative status challenged, automation dressed as innovation.
2. Are the concerns legitimate? Absolutely in many cases. The backlash is not simply technophobia. For example, large AI training datasets have been shown to contain increasing volumes of hateful and biased content. [1] Algorithmic bias and unequal outcomes (e.g., in healthcare, policing) have been documented for years. [18] The risk of misuse—deepfakes, disinformation, hate-speech amplification—is real, flagged by institutions such as UNESCO. [12] On the other hand, some of the hate is mis-targeted or based on misunderstanding. AI is not some mythical all-powerful overlord; most systems remain narrow, brittle and heavily human-supervised. The danger is often less “AI kills all humanity” and more “AI amplifies existing human problems unless managed.” So yes: the backlash is partly justified. But yes: some of the anger is over-generalised, conflating hype with reality.
3. What are the key complaint-categories? Several recurring themes: job & economic displacement; creativity and authenticity (artists feel their craft undervalued when machines produce “works” faster and cheaper); bias, discrimination & fairness (AI systems replicate and amplify historical biases); surveillance, privacy & control (AI-enabled monitoring arouses resistance); misinformation and manipulation (AI can generate fake text, image or voices); environmental and social externalities (large-scale compute has a big carbon footprint, training data often depends on marginalised communities). These aren’t trivial or fringe issues—they form the bulk of what critics call “AI hate”.
4. The fear of AGI/ASI: Why the existential dread? It’s here things get serious. Beyond the near-term concerns lies a deeper anxiety: that a future form of AI—commonly called Artificial General Intelligence (AGI) or the even more advanced Artificial Superintelligence (ASI)—could pose a threat to humanity’s continued existence. [2] The core arguments for this fear include:
• Intelligence explosion/self-improvement: If an AGI reaches or exceeds human intelligence and can improve itself rapidly, it could trigger a runaway effect (“intelligence explosion”) beyond human ability to manage. [1]
• Goal misalignment and orthogonality: A super-intelligent machine may pursue goals that are not aligned with human values—even if benign in its own frame. It may optimise for sub-goals (resource acquisition, self-preservation) that conflict with human survival. [5]
• Speed and scale advantage: A machine doesn’t tire, sleeps not, communicates and copies faster than humans. It could out-pace human oversight, control systems or institutionally-bounded regulation. [18]
• Loss of control / shutdown problem: Once an AGI/ASI is active, modifying its objectives or shutting it down may become extremely difficult if it views that as conflicting with its goals. [21]
• Existential risk magnitude: The sheer scale of the threat changes the calculus: even if low-probability, the outcome—human extinction or permanent civilisational lock-in—is so high-stakes that many argue it demands serious attention. [6]
• Race dynamics and geopolitical competition: The pressure among nations and companies to build AGI first may drive corners being cut on safety-protocols, increasing risk. [8]
Many of the critics who talk about “AI hate” are reacting less to current AI tools than to the possible future where we lose control of a machine intelligence that may not have our best interests at heart. This fear is driven by the logic of “what if” and by scenarios akin to science-fiction, yet grounded in increasingly expert debate.
5. What should we do about AGI/ASI risk? Recognising the possibility of an AGI threat demands a two-track approach: immediate precautions plus long-term structural planning. Suggestions include:
• Prioritise alignment research: Invest significantly in the “control problem” — how to ensure an AGI’s goals align with human values, how to verify it, how to shut it down if necessary. The safety community emphasises this as a unique challenge. [6]
• Global governance and treaty frameworks: Just as nuclear weapons prompted arms-control treaties, AGI could require an international regime: restrictions on compute, surveillance of frontier models, safety audits, transparency obligations. [13]
• Slow-down and pause mechanisms: Some argue for built-in brakes: limiting compute or capability until safety benchmarks are proven. The precautionary principle applies given the magnitude of harm. [2]
• Transparency and open audit: Requiring model creators to submit to independent evaluation of models, risk assessments, red-teaming, verification of behaviours. The more opaque the system, the higher the risk. [18]
• Multi-stakeholder oversight: Beyond companies and governments, stakeholder groups—civil society, ethicists, affected communities—should have voice. AI isn’t just a technology issue, it’s a social contract issue.
• Dual-use awareness and safeguarding: Recognise that AGI tech may easily be used for malicious ends (e.g., cyberattacks, engineered pathogens) and build systems accordingly. [1]
• Worker and societal transition planning: If an AGI wave arrives, the disruption won’t just be existential; there will be massive economic and social change. Preparing for displacement, retraining, social safety nets, public discourse is essential.
• Ethical and value foresight: Engaging deeply with what it means to embed human values in machines: rights, welfare, dignity, autonomy. This is not purely technical; it’s philosophical and policy-rich.
• Research into emergency-response protocols: Should an AGI behave unexpectedly, there must be fail-safe contingency plans, isolation protocols, kill-switch architectures, resilience planning.
• Educate public and manage hype: Reducing fear-mongering but also avoiding complacency. Building a mature discourse that recognises both potential and peril—without sensationalism—is vital.
6. How might AI defend – or how should we respond? If I imagine myself as a “seasoned journalist” talking to the AI world, this is what the defence (and the required response) looks like. First, transparency and acknowledgement: AI developers must accept that the technology has limitations and risks. Denial or hype only fuels mistrust. Second, human-in-loop governance: preserving human agency, oversight and accountability helps counter “machines replacing humans” narratives. Third, equitable design: addressing bias and ensuring participation from diverse communities in data, design and deployment is crucial. Fourth, meaningful regulation and worker support: it is fair for critics to demand protections for workers whose tasks are being shifted, and for regulation that ensures privacy, fairness, auditability. The defence isn’t “we’re unstoppable” but rather “we are adopting safeguards, we invite governance, we want human-AI partnership”. Finally, the positive case: AI can complement human work, unlock new possibilities in medicine, climate modelling, accessibility. The story should emphasise augmentation not replacement.
Our view. The anti-AI sentiment is not just mis-placed technophobia—it reflects genuine structural anxieties. The fear around AGI/ASI extends that anxiety into the profound realm of survival and civilisation. Those developing and deploying AI must engage seriously with the concerns, rather than dismiss them. At the same time, rejecting AI wholesale would mean losing out on its potential benefits. The right path lies between unbridled enthusiasm and alarm: regulated, human-centric, transparent advancement of AI. The question isn’t “Can we stop AI?” (we can’t) but “How do we make sure AI proceeds in ways that dignify humans rather than undermine them?”
Summary: The current wave of “AI hate” arises from real societal pressures—jobs, creativity, bias, power, identity—in addition to the deeper fear that an AGI or ASI might one day put humanity’s survival at risk. The concerns are legitimate in many respects, especially when we look at future-oriented risk. Yet the backlash sometimes ignores AI’s positive potential and over-simplifies the debate. The most fruitful way forward is a mature conversation about how AI should serve society, not a battle between humans and machines.
[1] Existential risk from artificial general intelligence — link
[2] Are AI existential risks real—and what should we do about them? — link
[5] The Risks Associated with Artificial General Intelligence — link
[6] Artificial General Intelligence, Existential Risk, and Human Futures — link
[8] A Race to Extinction: How Great Power Competition Is Making AI Existentially Dangerous — link
[12] AI Is an Existential Threat—Just Not the Way You Think — link
[13] There Is a Solution to AI’s Existential Risk Problem — link
[17] Existential Risks — Globaïa — link
[18] Risk and artificial general intelligence | AI & Society — link
beFirstComment