Revenge of the meat bags

• A small but widening circle of big-tech critics, artists, philosophers, humanists — and some technologists, too — are devising their own playbooks for a human response to AI's rise.
• These strategies for creating AI-minimal preserves in our public and private lives are bubbling up in corners of the old web and the new AI underground.
• Think of them as, at best, inspirational jumping off points for creative thinking, and at worst, rearguard actions in a fight for species survival.

1. Cool AI's jets

Today's furious AI race is driven by market competition and global rivalries — but the AI industry itself once charted a different, more cautious approach.

• Startups and tech giants alike were long committed to rolling out new AI work with great care to protect society from short- and long-term harms.

All that changed when ChatGPT's debut kicked off the generative AI frenzy at the end of 2022. A 2023 push for a six-month "pause" in AI work to weigh existential risks won only headlines.

• But a wide range of AI skeptics still hold out hope today for slowing the pace of AI development — either through government regulation or market discipline in a downturn — to make more room for human concerns.
• "Slow AI" is also a banner being waved by high-profile AI critics and ethics-first designers.

2. Slow ourselves down

AI's great promise is to boost productivity by speeding up task completion in the workplace. If that's not your goal, the technology becomes a lot less attractive.

• For technology users who already feel addicted to social media and tied to their smartphones, AI can look like a new, deeper circle of screen-overload hell.

Some AI refuseniks are rediscovering the "slow web" movement of at least a decade ago that rose as a response to social media platforms' online dominance.

• They're prioritizing human interactions over bot dialogue.
• Instead of surfing endless feeds, they're finishing a handful of longer articles (and even books!). They're taking plenty of exercise breaks — and also, when needed, naps.
• The artist and writer Jenny Odell — author of "Saving Time" and "How to Do Nothing" — is another inspiration for this approach. So is Kristoffer Tjalve's Naive Weekly newsletter.

3. Embrace human-made media

The media world, which fears losing traffic and revenue to a chatbot-dominated internet, is cycling back toward emphasizing creators' personal voices.

• The rise of audience-supported, individually produced email newsletters and podcasts suggests a continued hunger for imperfect, human-generated news, opinion and conversation.
• Reddit, one of the last redoubts of the once-teeming web of user-generated forums, now bills itself as "the human internet" — as opposed to the one that's crawling with bots.
• Wikipedia, whose articles make up some of the most valuable training data for many AI models, relies on a slow process of human review to fine-tune its "neutral point of view."

These are old moves in the digital media world, but AI is giving them new urgency.

• Every era redefines authenticity, and today "authentic" is coming to mean anything that's not AI.

It's not a good sign for AI makers that AI-generated material is now widely called "slop."

• In the AI age, we may quickly tire of LLMs' homogeneity — no matter how fine-tuned the personalization — and seek anything that feels real to us and can help us feel real, too.

4. Get "spiky"

To stand out in a world of bots, one effective maneuver is to be "spiky" — to act unpredictably, improbably, taking the paths that are least likely to be the consensus choice of an AI model trained on everyone's preferences.

• Students at an AI-first Austin school recently profiled in the New York Times call this the "spiky point of view."
• "To be a useful person in the age of AI, you have to have unique insights that AI doesn't really agree with," Alex Mathew, 16, a rising senior at Alpha High School, told the Times.

Yes, but: While "spiky" behavior could mean greater originality and creativity online, antisocial behavior, nihilism and random destructive acts meet the definition as well.

5. Stay offline

Every human act of spikiness will become training data for tomorrow's AI models.

• That will annoy the people involved — and also kick off escalating cycles of weirdness in AI output.
• This catch-22 is already driving some creators toward the spikiest act of all: keeping their ideas entirely offline.

Robin Sloan — the science fiction/fantasy author who Wired recently dubbed "The Tech World's Greatest Living Novelist" — is one proponent of this move.

• Sloan is no technophobe; he has experimented with AI for years. But he has lately taken to publishing his thoughts on printed broadsheet zines that have no "web versions."

In one of these from May, titled "The Secret Playbook," Sloan offers creators a treasury of zigs to counter AI's zag.

• For instance: "How do you mystify the AI model, frustrate it totally? Keep things offline. The great artistic movement of the 21st century will exist only in whispers between people."

Other Sloan advice: "Be stubborn," a quality AI models never display because "it is the one behavior that users will not tolerate." Also, "have fun," because AI models can't.

The bottom line: AI utopians who anticipate medical breakthroughs, universal education and economic abundance will argue that "keeping things human" is beside the point. For everyone else, the broad infusion of AI in our lives could trigger a strong recoil.

This is second in a two-part series. Read part one on the bot vs. bot future..