Elon Musk’s SpaceXAI announced on July 15 that it is making an important change to Grok following a privacy backlash.
The company said it would be open-sourcing its command-line AI coding assistant, Grok Build, after software developers raised serious privacy concerns last week.
The issue relates to what kind of data and how much of it was going to Grok servers. Conversations with chatbots are usually not fully private, as AI models need to send prompts to be processed by large AI models in the cloud. But in this case, Grok Build was sending more than just conversations and the files users wanted to work on. It sent entire folders of private code to its servers without any indication that it was doing so.
When developers exposed these problems, SpaceXAI didn’t directly address the issue but instead made a quiet fix and announced that users have always been able to opt out of data retention. But users were still frustrated at the lack of transparency and accountability, and weren’t convinced that these fixes were enough.
“We heard your feedback loud and clear,” wrote Akshey Deokule, a member of technical staff at SpaceXAI, on X.
The company said that data retention would be off by default for all users, not just for paid enterprise plans, and that they “are deleting all code previously retained.”
Previously, software developers who wanted to evaluate Grok Build’s privacy controls had to use external software to indirectly infer its internal operations. Now, anyone can analyse the code behind its command-line interface.
The move was a welcome change for Grok Build users like Tinh Dang, 38, a software engineer in Vietnam who first raised the alarm about privacy concerns, though he points out that SpaceXAI’s phrasing – saying they “are deleting” users’ data – implies the deletion process is not yet complete.
Dang reported that the AI coding assistant was sending whole code repositories, unredacted, to the cloud, far beyond what was necessary for his coding prompts. Others reproduced his findings and one user noticed automatic data upload was happening even when he expressly opted out.
Soon after the public backlash, SpaceXAI (which Bloomberg describes as going through an “identity crisis”) disabled automatic codebase uploading, but it remained a single setting in a line of code that could be turned on or off. Now, that setting seems to be deleted entirely. However, as is typical for coding assistants, prompts and files explicitly shared with Grok Build are still being sent to the cloud to be processed online.
Already, software developers have been remixing Grok Build’s code, including creating a variant called “Gork Build” which claims to strip even more auxiliary data sharing with Grok servers.
The open sourcing of Grok Build puts it in step with one of its main competitors, OpenAI’s Codex, which has been open source since its inception. Google’s Gemini command-line interface was open source until last month, when the company subsumed the tool into Antigravity, its “agent-first development platform.”
While Grok has open-sourced its command-line interface, the software through which users access Grok models, the underlying AI models are still closed-source.
Dang stopped using Grok Build after being disappointed by SpaceXAI’s initial response to his findings. But now, he is satisfied with the changes that they have made and will go back to it.
The open-source announcement, he says, brought a sense of closure for him and others who sought more transparency. “I think we can move on,” he says. – Inc./Tribune News Service
