Cloudflare, a US-based company that uses its global network and cutting-edge tools to enhance website speed, safety, and reliability, has implemented AI oversight features in Cloudflare One, its enterprise security platform.

These new features will enable IT departments to quickly identify who uses AI and what types of data they share with the technology. According to Cloudflare, this is a reliable way for the IT workforce to get an “X-ray vision” of how employees are using generative AI, which is also already built into the available dashboard.

This came after research from sources noted that several employers in different institutions embraced the introduction of generative AI in their companies’ operations. However, certain employees share internal financial statements or proprietary code with AI chatbots such as ChatGPT, Claude, or Gemini, which is getting exposed.

The revelation has caused employers to lose confidence in the technology. Sharing financial details with AI chatbots exposes the firm’s secrets, ending up in the cloud.

In a blog post, the company highlighted, “Admins can now find answers to questions such as: What are our employees doing in ChatGPT? What kind of information is being uploaded and used in Claude? Is Gemini set up correctly in Google Workspace?”

Cloudflare cautions firms on reckless actions while handling sensitive information 

After conducting thorough research on using AI chatbots, Cloudflare discovered that three out of four employees include ChatGPT, Claude, or Gemini in their daily operations. This is particularly for tasks such as evaluating data, editing text, debugging, and designing. 

Considering these practices, the company identified a primary problem that may arise when using AI. They mentioned that companies’ sensitive details may get uploaded to these AI chatbots untraceably. To address this, the US firm’s products are designed to work at the API level and monitor any suspicious uploads in the system.

Now that such a scenario has been discovered, the company has cautioned individuals to avoid reckless actions while handling sensitive information. Based on Cloudflare, careless actions may lead to an external model training using the shared confidential details, and these details could later disappear forever.

Other major firms in the field of enterprise security, like Zscaler and Palo Alto Networks, also provide AI oversight. Even with these firms in place, Cloudflare still claims that Cloudflare One is unbeatable, arguing that it has a hybrid model that works perfectly without installing software on devices. The platform also features API scanning to monitor system posture, configuration issues, and data exposure, while providing instant control over prompts across ChatGPT, Claude, and Gemini.

In the meantime, it is worth noting that Cloudflare has established itself as a content-neutral infrastructure provider instead of a moderator. To break this down, the firm only monitors what its clients publish when they are authorized to do so. This approach has been implemented for the past decade. 

Matthew Prince, the CEO of CloudFlare, weighed in on the topic of discussion. According to Prince, the company is not a hosting service and has no authority to choose what content is allowed. However, he stated that its main focus is ensuring websites are up and running regardless of their views.

Cloudflare faces criticism from individuals on its website services

Cloudflare’s free speech stance has drawn criticism, with opponents arguing that the company enables harmful websites. This includes those with hate or extreme content. This, they say, is because no formal request to delist has been submitted.

These allegations followed a 2022 study by Stanford that pointed out that Cloudflare is responsible for hosting some of the top misinformation websites, compared to its total internet traffic.

Despite this, several exceptions have been made for the company. For instance, in 2017, the firm quit offering services for the white supremacist site The Daily Stormer just after it reportedly claimed that Cloudflare offered support for its pro-Nazi stance.

A short while later, Prince qualified their announcement as an unwilling exception, which had been wrested from them and forced them to break their neutral policy.

Don’t just read crypto news. Understand it. Subscribe to our newsletter. It’s free.