Google has restricted access for several users, including subscribers to its Antigravity and Gemini AI Ultra services, after detecting what it described as violations of its terms of service related to use of third-party AI tools.
The restrictions follow reports from affected users that their access to advanced models such as Gemini 2.5 Pro — and in some cases linked services including Gmail and Google Workspace — was abruptly blocked without warning.
Varun Mohan, a former CEO of Windsurf and engineer with Google DeepMind, said a “massive increase” in malicious activity on the backend of Google’s AI coding platform Antigravity prompted the enforcement action.
“We’ve been seeing a massive increase in malicious usage of the Antigravity backend that has tremendously degraded the quality of service for our users,” Mohan wrote in a post on social media platform X on Monday, February 23.
Mohan said Google detected that some users were accessing protected models via third-party tools such as OpenClaw, an open-source AI agent framework that can automate tasks such as email handling, flight check-ins and insurer interactions — activities that potentially conflict with Google’s terms of service.
“We needed to find a path to quickly shut off access to these users that are not using the product as intended,” Mohan said. He noted that some users may have been unaware that their actions violated Google’s ToS and suggested the company would provide a route for compliant users to regain access.
OpenClaw drew rapid attention after its November 2025 release, amassing more than 100,000 stars on software repository GitHub and drawing millions of visitors. However, agentic AI tools like OpenClaw have also attracted scrutiny, including warnings from China’s industry ministry that they could pose security risks when improperly configured.
In a similar move last week, AI developer Anthropic updated its consumer terms of service to explicitly ban the use of OAuth tokens obtained through Claude Free, Pro or Max accounts in third-party tools, including agent SDKs.
Commenting on Google’s approach, Peter Steinberger, creator of the OpenClaw framework, called the restrictions “draconian” and said he will remove support for Antigravity. He contrasted Google’s stance with Anthropic’s more collaborative approach to addressing compliance issues.
Earlier this month, OpenAI CEO Sam Altman announced that Steinberger would join the ChatGPT maker to “drive the next generation of personal agents,” with OpenClaw living on as an open-source project under a foundation supported by the company.
The account restrictions highlight growing tensions in the AI space as major providers tighten controls over how their models are accessed and integrated into third-party systems, even as developers push the boundaries of agent-based automation.
