Rabbit R1 security vulnerability allows third parties to view private AI responses

A rather serious security vulnerability has been uncovered on the Rabbit R1, one of the latest hardware devices built by an AI-first team. The flaw allows third parties to access text prompts sent through the R1, which can contain any amount of sensitive info.

The Rabbit R1 came as another product developed by a company that “specializes” in AI development. It’s essentially a handheld device that allows you to prompt or ask it to accomplish certain tasks. Those tasks might include “order me a pizza” or “let my partner know I’ll be late.” It isn’t a phone replacement, but it’s pretty close – at least, it was meant to be. Those requests are supposed to be secure transmissions to Rabbit’s “rabbithole,” a cloud-based processing system that takes AI queries and turns them into action through your connected applications.

It turns out, though, those transmissions are not as secure as one may think. According to a report from rabbitude – a community of Rabbit R1 developers – requests sent can be retrieved via hardcoded API keys (via Engadget). The report notes that these keys contain responses the R1 devices respond with.

Update 6/28: Rabbit has since informed us that these API keys have been rotated, noting that the company has begun taking action to revoke and ensure the “secrets,” or API keys, are properly stored. The company did not, however, comment on the notion that these keys, no matter what entity’s hands they lie in, still contain R1 responses with private data. You can find the company’s security page through its website.

The inherent issue here is that these responses often contain sensitive information. That information can then possibly be made available to malicious actors who have access to the hardcode keys. Some of the apps referenced in the report are Google Maps, Yelp, Azure, and more. It’s easy to see how some of these responses can contain information that one wouldn’t want available without a significant amount of effort.

Rabbitude stated that Rabbit knew about the R1 security issue and had not taken action to rectify it. Rabbit, in turn, told Engadget that an “alleged data breach” came to its attention, but nothing was found indicating a vulnerability in the system.

As of right now, we are not aware of any customer data being leaked or any compromise to our systems. If we learn of any other relevant information, we will provide an update once we have more details.

Rabbit

The Rabbit R1 isn’t a stranger to criticism. When the device released, it quickly became clear that the device was not ready for prime time, even if was only a $200 smart assistant.