Copyright in the Age of AI

October 19, 2024 — by Per Christensson

ChatGPT is an amazing technology, but it is killing the web.

For several decades, people Googled their questions. They typed in search terms, viewed the search results, then clicked on a website. Visitors allowed website owners to monetize their original content through ads. The more traffic they received, the more revenue they would generate, and the more content they could produce.

With ChatGPT, you can ask a question, get an answer, and have no idea where the answer came from. The information you receive is aggregated from copyrighted sources — mostly websites — often with no citations.

While generative AI is tremendously helpful for users, it's destroying the livelihoods of content creators. Imagine thousands of people using your original content on a daily basis but never visiting your website. You don't see any website traffic or make any revenue from ads. What is the incentive to keep creating content?

Over the past several years, I and other webmasters have seen our website traffic drop substantially. I've seen a lot of sites stop producing content altogether. Not only does the trend seem to be getting worse, it appears to be accelerating.

What's Being Done

In December last year, the New York Times sued OpenAI for scraping their content and reusing it without their permission. It was a landmark lawsuit, but it was almost a year ago, and no substantial action has been taken. Meanwhile, ChatGPT has exploded in popularity, and hundreds of millions of people are now using it on a daily basis.

More promising is Cloudflare's initiative to charge AI bots for scraping websites. The global CDN is already collecting scraping data and will soon provide a way to charge bots for scraping (and using) website content.

▶ Cloudflare's effort to help webmasters is noble, but the company may be the biggest beneficiary if its idea is successful. Even a tiny fraction of the revenue across their global network of sites would be substantial.

Many questions remain, such as how much will AI platforms pay content creators? Will they get paid for each scrape or only for new content? Will webmasters only earn money if the services use their content?

Most importantly, will there be enough incentive for people to publish original content? If not, AI may destroy the web as we know it. We'll be left with AI-generated content created from other AI content. Even the bots will get tired of that.