Companies are increasingly limiting access to AI tools: millions of DNS queries blocked

Companies are increasingly disrupting traffic to generative artificial intelligence (AI) tools so that their employees cannot reach Notion, ChatGPT, Claude, or other assistants. DNSFilter, a protected Domain Name System (DNS) service, said it has already blocked 60 million visits.

DNS acts like a phone book for the Internet, translating website addresses into their corresponding IP addresses. Private DNS services offer domain name-level protection, filtering out unwanted domains and allowing companies to choose what they want to limit.

While the services are mostly used for protection against malware, trackers, or other dangerous domains, some companies are increasingly using DNS tools to block access to AI websites.

DNSFilter said it offers a “generative AI category” that allows organizations to block and filter these domains with a single switch.

“In March, the company blocked over 60 million generative AI requests, representing about 12% of all such queries,” the company said in a press release.

The actual proportion is likely higher due to the way the DNS system works. If an employee fails to reach the AI service once or twice, they are unlikely to repeatedly attempt to access the website.

A frequent AI user may generate tens, hundreds, or even thousands of DNS queries with each visit, depending on the TTL (Time To Live – the expiration term for the DNS record, usually in seconds or minutes).

“Companies know the benefits that generative AI offers, but they also know the potential cybersecurity risks. More organizations are now proactively choosing which tools to block and which to allow,” said Ken Carnesi, CEO and co-founder of DNSFilter.

“That way, employees can't 'sneak' AI tools into the corporate network or inadvertently use a malicious one. In this way, a DNS filtering solution helps companies enforce policies, block possible threats, and enable greater productivity all at the same time.”

Since January 2024, DNSFilter has been processing an average of over 330 million monthly queries that fall under this category.

“Notion was the most blocked of all generative AI queries in March, accounting for 93% of all blocked generative AI queries – and far more than the number of blocked queries for Microsoft Copilot, SwishApps, Quillbot, and OpenAI combined,” DNSFilter said.

The firm explains that companies are becoming more selective about which domains they allow employees to use. Internal policies typically aim to reduce shadow AI by directing employees toward approved tools based on clearly defined guidelines.

“For some organizations, blocking certain domains can be an effective way to set policy, protect the company, and deploy more control of what tools are being used.”

DNSFilter also sees a trend of cybercriminals creating malicious generative AI domains, often mimicking well-known companies.

Between April 2024 and April 2025, there was a 2,000% rise in malicious sites with “openai” in the name.