135: The scary truth about AI for publishers

The challlenge of AI
Summary: The article discusses the challenge that artificial intelligence (AI) poses to publishers, as AI can provide answers and content similar to what traditional publishers offer. The author suggests steps for publishers to respond, including denying AI access, protecting their content, and characterizing AI answers as unreliable. It also addresses the issue of users uploading content to AI chatbots. The author acknowledges the difficulty in preventing this legally and suggests it may require lawsuits, regulations, or legislation.

The challenge that artificial intelligence poses for publishers is pretty simple. If I can get an answer from ChatGPT, why do I need to go to your site, or read your magazine? In a year or two, the idea that I would enter a search and be content with a list of homework to do to find the answer to my question will seem as outdated as the stirrup cup.

The fact that AI can provide answers to questions is only a threat to publishers who provide answers. If you provide comedy, or drama, cat videos, or something like that, …. Oh, wait. AI does that as well. Nevermind.

It’s a challenge for all publishers.

In response, I’ve encouraged all publishers to take four steps.

  1. Change your robots.txt file to politely ask the AI bots not to slurp up your content.
  2. Since that’s not good enough, also get your IT people to deny such requests at the server level.
  3. Change your terms and conditions to make it clear that your content may not be used to train AI models.
  4. Start training your own AI models with your own content and your own voice so you also can answer people’s questions.

I have two new things to add to this list.

  1. Start characterizing the answers from AI as misinformed, erroneous, unprofessional, untrustworthy, etc. If publishers follow my advice, and if we’re lucky, that will turn out to be true as informed, professional, trustworthy content will no longer be available to train their models. They’ll be stuck with tin-foil hat blogs and such.
  2. Further amend your terms and conditions to say that users may not upload your content to AI chatbots.

Unfortunately, that latter point is likely to be a losing battle.

Consider this. You hire professional editors to carefully research and explain the latest development in whatever subject you’re expert in. It’s great stuff, and ChatGPT couldn’t hold a candle to it.

Then one of your subscribers uploads your report to ChatGPT and says “please summarize this for me.” There are even browser plugins that will do this automatically, so that all your carefully researched content is fed to the AI monster.

I don’t think there’s a technological way to keep your content from getting into the maw of the AI chatbots, so you’ll have to try to stop it legally in your terms and conditions. And I guess in lawsuits, regulations, and legislation.

Sorry. I don’t like it either, but when I look into my crystal ball, that’s what I’m seeing.

Or you can just roll over and give up. I suppose you could do that as well. But I’d rather you didn’t, because then I wouldn’t have any clients.

Leave a Reply

Your email address will not be published. Required fields are marked *