Key Takeaways

  • Anti-AI sentiment isn't a single wave but a "big fuzzy mess" of real grievances, economic anxieties (like electricity bills), cultural clashes (the "AI slop" debate), and outright misinformation, according to tech analyst Benedict Evans.
  • Specific alarmist claims, such as data centers consuming massive amounts of water, are false. Evans notes a Livermore Lab study in late 2024 estimated US data center water use at only 0.017% of total national consumption.
  • Despite widespread fear, current economic data shows “no clear consensus that we're seeing an impact on jobs” in the economy right now, challenging public perceptions of mass unemployment.
  • The current backlash against AI mirrors the internet's early days, where the power to connect people for good (finding your "tribe") also enabled negative forces (like finding "bad people" or pedophiles), a lesson Evans says will repeat with AI.
  • A growing "culture war" around AI-generated content, highlighted by estimates that 30-40% of new podcasts are AI-generated, feeds distrust and public skepticism about authenticity and quality.

The Fuzzy Mess of AI Backlash

For ambitious founders, the anti-AI sentiment feels like a growing headwind. But what exactly is it? Benedict Evans calls it “a big sort of fuzzy mess of different stuff.” It's not one unified critique, but a cocktail of everything from legitimate worries about deepfakes to anxieties over energy bills and job security, all swirled with outright fiction. Evans points to cultural skirmishes, like the "AI slop" debate, where creators push back against synthetic content, noting "30, 40% of new podcasts generated by AI." This isn't just noise; it’s a sign that trust, or the lack thereof, is becoming a critical component of market acceptance.

Debunking the Water Myth (and the Job Panic)

Amidst the fuzzy mess, some specific claims gain outsized traction. One popular narrative suggests AI data centers are sucking up our planet's water supply. Evans took the time to dig into this. He found that, objectively, the numbers don't support the panic. “The water thing is weird because it's just like completely fake,” he states. A Livermore Lab study, published near the end of 2024, estimated US data center water consumption at a tiny 0.017% of total national water use. That’s less than one-fiftieth of one percent. Founders need to recognize that some criticisms are based on easily disproven falsehoods, and prepare to counter them with data. Similarly, while jobs are a constant concern, Evans highlights, “there's no clear consensus that we're seeing an impact on jobs” in the economy today, despite public fears. It’s a fear-driven narrative, not yet an economic reality.

Echoes of Social Media's Dark Side

For founders building the next generation of AI tools, Evans offers a potent historical parallel: the early internet and social media. In the 90s, the web was hailed as a connector of communities. “You can be...the only gay kid in your village and you can find other gay people and you can find your tribe,” Evans recalls the optimism. But that connection had a dark underbelly: “it turned out you could also be the only Nazi in your village or the only pedophile.” The technology didn't just connect the good; it amplified “all the bad people and all of our own worst instincts and every problem in society.” Evans' blunt assessment: “that will happen again with AI.” This isn't a cynical prediction; it's a warning. Expect unintended consequences, societal fractures, and the amplification of humanity's darker impulses alongside its brightest innovations.

What to Do With This

Founders, stop treating anti-AI sentiment as a PR problem to spin. Instead, treat it as a critical market signal. When designing your next product, don't just ask what AI can do; ask what responsible, transparent AI should do, specifically anticipating its potential for misuse and unintended societal friction. Arm your sales and marketing teams with hard facts—like the negligible data center water consumption—to disarm misinformation early and build trust where none exists. Proactively design features that build transparency around AI's role and mitigate "slop" or deepfake potential, rather than waiting for regulatory pressure or a user revolt.