Show me the incentive, and I’ll show you the outcome — Charlie Munger
Uber was heavily lobbied against by various taxi interests as the service grew. Disruption isn’t a business model in itself, but it’s usually an interesting enough phenomenon to garner further inspection. So why not look at some of the heaviest lobbyists against generative AI to see where it might be most useful?
- StackOverflow banned ChatGPT or any LLM when posting content (announcement).
- Nikon ran an anti-AI advertisement, “natural intelligence,” “don’t give up on the real world.” (ad)
- Away (the luggage company) also ran a similar ad to Nikon’s “the world is not AI-generated” (ad)
- The Red Cross in France also ran a similar “this photo was not created by artificial intelligence” (ad)
- The Directors Guild of America reached a three-year agreement with the Alliance of Motion Picture and Television Producers to safeguard their members against AI advancements (detail).
- The Writers Guild of America has been on strike since May with the hope to limit the use of AI.
- New York Public Schools and the Los Angeles Unified School District (both K-12) have banned ChatGPT, citing that “the district does not allow cheating and requires original thought and work from students.”
- ArtStation, a portfolio site that was included in many of the training corpora for AI models, hides AI-generated art from its homepage, claiming the content violated its Terms of Service. (article)
- Valve refuses to distribute games with AI artwork that can’t prove they have the legal rights to all their training data (article)
- The journal Nature banned AI-generated art (announcement)
- Google stated that content automatically generated with AI writing tools is considered spam (article).
- Getty Images banned AI-generated content on the basis of legal challenges (article)
- A Texas judge banned legal filings that are drafted primarily by AI (article)
In all of these cases, there is
- Someone already using generative AI to automate or augment a task
- A challenge to the system — shifting responsibility, automation of a job, or disrupting supply and demand.
Some of these anti-AI decisions might be long-term directionally right, but some of them might be short-term reactions to future disruptions. A counter-intuitive learning in startups is that when things work, they usually work quickly.
On the other hand, groups that are also thought to be disrupted or obsoleted by AI aren’t protesting but adopting it in record numbers. How many programmers have rejected AI tools? GitHub Copilot is the most successful enterprise deployment of LLMs to date. Marketers have signed up for AI copywriting services and fueled a positive feedback cycle in the VC funding. Mathematicians like Terrance Tao have embraced LLMs rather than decried them.