What do you need to know
- Cloudflare is a global cloud provider that provides security and DDoS protection to millions of websites and protects roughly 20% of all global internet traffic.
- Yesterday, Cloudflare announced a new free tool it will begin offering to all customers specifically designed to wipe out AI crawlers.
- AI bots used by companies like Google, Microsoft, OpenAI and many others steal copyrighted information from websites like ours to train their premium AI services.
- Last week, Microsoft’s AI chief said all publicly available internet content was “freeware” to be stolen for its AI aspirations.
Last week, Microsoft’s AI chief said that public content on the open web is “freeware,” giving the trillion-dollar company a free license to steal any content you’ve published on the web to power its premium products. The backlash against the mess was significant and served as a kind of clarion call for web content providers to rethink their relationship with companies like Microsoft, which seek to profit from the hard work of content creators and give literally nothing in return. Cloudflare may have just handed those same creators a very important defensive weapon in the fight back.
Cloudflare is a global internet service and hosting company that powers roughly 20% of all web traffic. Offering things like DDoS attack protection and website bot validation checks, Cloudflare helps improve the general quality of the world wide web and uses its massive server infrastructure as an extensive layer of security for companies of all shapes and sizes.
Yesterday, the company announced a new feature it will begin rolling out to all users, even those on its free tier, designed to combat generative artificial intelligence.
To help keep the internet safe for content creators, we’ve launched a brand new “easy button” that blocks all AI bots. It is available to all customers, including those on our free tier. See our blog post for more details: https://t.co/csWFFgqbKMJuly 3, 2024
Declare “AIndependence,” says Cloudflare on its blog. Its new system will allow users to opt-in to block AI bots and crawlers from accessing websites, effectively preventing Microsoft, Google, OpenAI and others from stealing web content for free.
After surveying its users, Cloudflare shared data that more than 80% of its customers wanted the ability to prevent Microsoft from stealing their content. “We’ve heard clearly that customers don’t want AI bots visiting their sites, and especially those that do so dishonestly.” Cloudflare continued: “To help, we’ve added a brand new one-click to block all AI bots. It’s available to all customers , including those on the free tier.”
Generative AI training content is becoming lucrative and valuable for companies like Google and Microsoft. Google reportedly paid over $60 million to access all of reddit’s content to train its models, which also laughably led to sarcasm and trolling appearing in Google search results.
Can Microsoft find a healthy balance?
I’ve written before about how it should be in Google’s and Microsoft’s interest to create a healthy, symbiotic relationship between human creators and their generative AI efforts. Generative AI undoubtedly plays a role in the future of technology, but I feel like companies are still struggling with exactly what that looks like for customers. Right now, generative AI seems to be best used for the most basic writing tasks, like creating formal emails or summarizing a long piece of text. Even then, it has trouble doing even basic things when you mess with it. I’ve found that this often simply detracts from productivity rather than increasing it, given that you have to double-check everything the AI ​​does to avoid their “hallucinations”.
AI is also incredibly expensive to operate. AI queries are disrupting Google’s emissions reduction efforts, and I doubt Microsoft is doing much better. Even if you discount the climatological impacts, even the business model doesn’t seem to be working exactly today. Microsoft gives Copilot away for free and I’m not sure why I would ever pay for it.
RELATED: Why Microsoft won’t be the company that mainstreams AI with consumers
A low-hanging fruit feature that Google and Microsoft have been very quick to pick up on is search summaries. We produce thousands of guides here at Windows Central and Microsoft Copilot now simply takes the content of the article and reproduces it, depriving us of traffic and therefore income. This is bad for us, but also bad for Microsoft and Google. If human content creators can no longer effectively monetize and make a living, more and more of the internet will be generated by artificial intelligence. Similar to JPEG compression, content quality will degrade once the AI ​​starts learning from other AIs, rather than from human creators. Because ultimately AI does not “understand” the content it reproduces and can only infer context by comparing it to human content. This phenomenon is called model collapse, and it’s a real problem among serious AI scientists. But right now, everything Google, Microsoft, and others think of is getting ahead.
This kind of technology still needs human intervention to really take off. The alarm raised by Microsoft’s AI chief over his irresponsible “freeware” comments has contributed to the ongoing backlash. And now companies like Cloudflare are stepping up to help fight back. It won’t be long before others follow, and Microsoft may really have to reckon with its cavalier approach to content theft on an industrial scale.