Resist "AI"
Critics of generative AI argue that its environmental costs, reliance on copyrighted training data, and potential to deskill workers and displace jobs outweigh its productivity gains. Supporters counter that AI is just another powerful tool—akin to past technological shifts—that can free people from drudge work and boost competitiveness, especially for those who embrace it thoughtfully rather than reject it outright. Underneath the technical arguments lie deeper disputes about capitalism, regulation, greed, and whether society can steer AI to serve human needs instead of exacerbating inequality and cultural decay.
Environmental & Resource Impacts
- Large AI/data centers’ water use is debated.
- Some argue evaporative cooling uses scarce fresh water, doesn’t fully return to usable sources, and worsens regional drought and climate stresses.
- Others claim global fresh water is abundant, scarcity is regional, and open-loop cooling is fine in water-rich areas; they see concern as overblown or misapplied.
- Critics see AI as another high-footprint industry (like power plants), deserving regulation on energy mix, water loops, and siting.
- Supporters counter that environmental harms are about current implementation, not AI per se, analogous to early dirty cars vs cleaner future tech.
Historical Analogies & “Inevitability”
- Many note past “anti-tech” reactions (writing, printing press, cars, Internet), arguing AI fears are structurally similar and that resistance leads to irrelevance.
- Others respond that some past tech critiques were valid, and “it’s like cars/printing press” is a shallow rebuttal.
- “Inevitable” framing is challenged: some compare AI restrictions to controls on nukes, guns, or harmful content.
Work, Productivity & Skills
- Pro-AI developers report major productivity gains, faster feature delivery, and higher income; they see AI primarily as a tool to automate drudgery.
- Skeptics say over-reliance erodes understanding of code and domains, especially when specs, tickets, and reviews are AI-written “slop” that’s verbose, low-signal, and hard to trust.
- Some fear domain expertise will vanish, leaving no humans capable of validating AI outputs. Others think new “AI orchestration” skills simply replace older low-level skills.
- There’s concern about burnout, lost craftsmanship, and jobs becoming less meaningful, especially for those who enjoyed “hand coding.”
Ethics, Copyright & Data Use
- One side calls models “trained on stolen code/art/books,” sees commercial use as unethical and tantamount to exploiting unpaid human creativity.
- Others argue training is analogous to human learning; outputs are transformative; copyright’s purpose is progress, not pure creator enrichment.
- Sampling in music and fair-use doctrines are invoked on both sides, with strong disagreement over whether LLM training is more like reading, like unlicensed sampling, or something new.
Social, Educational & Political Effects
- Concerns include students using AI to bypass learning, worsening long-term competence; AI enabling confident but wrong “DIY experts” in medicine/law; and accelerating inequality under profit-driven leadership.
- Some call for “resisting AI” as resistance to greed and premature deployment; others see that as fear-driven or futile, arguing adaptation and targeted regulation are more realistic.