Companies use AI to read private Slack and Teams messages

It is not a secret that work chats are not private, but AI is taking it to the next level.

As reported by CNBC, huge employers such as Walmart, T-Mobile, Chevron, AstraZeneca, Nestle, and Starbucks are using Aware AI software to analyze employees' messages.

Founded in 2017, Aware provides integrations with Slack, Teams, Google Drive, Zoom, and Meta workplaces. By extracting data from communication platforms, AI provides employers with feedback on campaigns and initiatives at the company, the overall mood of the employees, and flags potential violations of the company’s policies.

According to the company’s press release, Aware uses a variety of techniques to analyze and interpret heterogeneous digital conversations, including NLP and CV neural network modeling, to identify individual and organizational patterns of behavior.

Aware's models are trained on a proprietary dataset of tens of millions of conversations, which, according to the company, results in more accurate and representative insights than generic datasets.

When a new client joins the analytics tool, it takes about two weeks for Aware's AI to train on employee messages to understand the company's patterns and to recognize what's usual and what's not.

While the companies have control over preserving the privacy of the employees, in the cases that are marked as causing the biggest threat, employers can get access to the identity of the employees. Employers can define the policies, which, once violated, could be flagged by the AI and the employee's name revealed.

A double-edged sword

In its risk assessment report, Aware claims that the main issues encountered in the 6.6 billion real messages were employees sharing sensitive information in screenshots or written text and employees using inappropriate or offensive language. According to the report, 1 in 95 messages contained toxic speech.

The company also identifies that over 90% of all messages sent in collaboration platforms occur in private or restricted channels unreachable to managers. There, passwords, code, intellectual property, and other corporate-sensitive data could be stored indefinitely, available for anyone with workspace access to find.

However, critics see the widespread use of such AI-driven tools as a potential threat to employees’ privacy.