Security Flaws Exposed in AI Projects by Scale AI’s Public Google Docs

Scale AI is facing serious security concerns after leaked documents revealed that it uses public Google Docs for sensitive projects, including those for major clients like Google, Meta, and xAI. These documents, labeled “confidential,” are accessible to anyone with the link, exposing sensitive information about AI training and thousands of contractors, including private email addresses and performance evaluations. Cybersecurity experts warn that this practice increases vulnerability to social engineering attacks and other hacks. Despite no confirmed breaches, contractors have noted that the system is inadequately secured, leading to potential data misuse. In response to these issues, Scale AI has stated it is conducting an investigation and has disabled public sharing of documents. The concerns coincide with a significant investment from Meta, prompting clients like Google and OpenAI to pause their work with Scale, raising doubts about the company’s commitment to data security and client trust.

Source link