Telkom warns of data breach risks from public AI usage

Jakarta — Telecommunications company PT Telkom Indonesia has cautioned against the data breach risks posed by using publicly accessible artificial intelligence (AI) websites or applications such as ChatGPT, Gemini, and similar services to assist with work and business activities.

Komang Budi Aryasa, Executive General Manager for Digital Products at Telkom Indonesia, stated that it is crucial for companies to establish restrictions on employees’ use of AI tools.

“The key restriction is that internal company documents and requirements must not be queried through open applications like ChatGPT that are freely accessible to the general public. This is because any data uploaded potentially risks being leaked,” Komang said during a media briefing in Jakarta on Monday.

He acknowledged that whilst AI usage in the workplace is indeed very convenient and helpful for workers, caution must be exercised.

“When we upload a document, it is stored on their servers. When another person submits a similar prompt, our document could be delivered to that other person—potentially even to a business competitor or irresponsible individuals. That is how data breaches occur,” Komang explained.

For this reason, Komang emphasised that companies must establish firm rules governing AI use in the workplace.

“Companies must anticipate these risks. It is not easy, certainly. Companies now need top-down governance—they must issue clear policies with proper enforcement mechanisms,” he said.

“If assistance is needed, use the internal GPT system that we have already developed. The results are the same,” Komang added.