Samsung lets employees use ChatGPT again after secret data leak in 2023

OpenAI's ChatGPT AI model can be a useful tool for work but the security considerations are different for an average joe versus an employee at a multinational corporation with trade secrets. Samsung first gave its employees access to ChatGPT to assist with their work back in 2023 but multiple incidents of sensitive data leaks forced it to implement new security procedures, including limiting the capacity of entries to ChatGPT.

The threshold for ChatGPT use has now reportedly been lowered across Samsung Group companies. Employees were primarily confined to using Samsung's internal Gauss AI previously, but the earlier restrictions have now been relaxed to increase work efficiency.

ChatGPT access has been expanded across group companies with enhanced security guidelines. The battery-making affiliate Samsung SDI has allowed employees access to ChatGPT, so has Samsung Display and Samsung Electronics.

Separate security protocols and procedures are in place to ensure that employees can surely use the external AI model. It's mainly being used in departments that have to work with different AI technologies instead of departments that deal mainly with product development.

This should help prevent inadvertent leaks. Previous incidents involved an employee at the semiconductor division entering source code into ChatGPT. One employee had entered code into ChatGPT to understand yields and other relevant information.

That's data that Samsung wouldn't out to be exposed outside the company at all. Since AI models use input by users to expand their training data, this meant that a bit of Samsung's source code became part of ChatGPT's data sets.

Samsung may view ChatGPT as a useful tool to help improve employee productivity. The top boss, Jay Y. Lee, met with OpenAI CEO Sam Altman back in February this year, so evidently no bridges were burned due to the 2023 incident.