Can you stop AI stealing your work?

Kathleen wrote in to say she’s worried about AI stealing her work.

Join the queue.

The short answer? Don’t put anything into an AI model if you’re concerned about copyright, i.e, don’t talk to it, install it, or use it.

And you’ll probably need to kill most of your SaaS subscriptions and social media accounts too… because they’re all feeding AI.

Considering all the “Now with AI” hyperbole on every app and service, we may as well quit using the web.

AI companies don’t see anything wrong with training their models on other people’s works.

In Empire of AI, Karen Hao explains many AI researchers and developers see how an AI model trains itself as NO different from an artist or creative getting inspired by one of their peers.

The argument being….

We only have films like Apocalypse Now because director Francis Ford Coppola felt inspired by Joseph Conrad’s Heart of Darkness.

But, I’ve yet to meet the creative who consumed the sum total of their genre or industry and went on to create a multi-billion dollar company.

For most of us, avoiding AI altogether is unrealistic…. but some workarounds exist.

Let’s say you’re running a website or blog:

According to OpenAI, you can disallow GPTBot from crawling your website by adding User-“agent: GPTBot” and “Disallow: /.” to your site’s robots.txt file.

Raptive, the ad network I use, recommends blocking ad crawlers by adding the following code to the header of every page on a website .

Its WordPress ad plugin blocks AI crawlers by default. Raptive claims publishers like The New York Times, Wall Street Journal, Vox, and Reuters are blocking access by AI crawlers.

If you’re not on the Raptive ad network, add these snippets of code using a WordPress plugin like Yoast or RankMath.

Or just ask ChatGPT how to do it!

But blocking AI bots will only protect new content.

If your blog posts or content are more than a year or two old, OpenAI has probably trained its models on it. And anything posted on social media is being gobbled up by big tech companies working on their models.

OpenAI’s documentation also doesn’t say anything about how to exclude content on The Way Back Machine (the web’s archive), either.

Meanwhile, multiple sources, including Cloudflare and Techcrunch, have accused Perplexity of bypassing robots.txt rules. I asked Perplexity to say it ain’t true, and the AI agent’s answer doesn’t exactly fill me with confidence.

So pick a side.

In one corner, we have big publishers and media orgs filing lawsuits against AI companies to figure out how they can get paid. And in the other corner, we have AI companies like Antrophic claiming a mass copyright lawsuit in the US could ruin the generative AI industry.

I’m not involved in any class action lawsuits, even though AI trained itself on my sites without permission.

I’m more interested in what I can do with AI than in wishing it were 2016 again.

I was also curious to read Cloudflare is also testing a pay crawl service. In private beta, this enables publishers to earn residuals if AI agents access your content.

I’d love to see a solution like this work so creatives get paid and use AI. But I don’t worry too much about AI stealing my job after the fact.

And AI, as helpful as it is, can’t replace human creativity.

Now, if you want to avoid AI, so be it. It’s not for everyone. I get that.

But if you’d rather figure out how to use AI for your business and combine it with your best ideas, check out Prompt Writing Studio.