Samsung’s chip division is at a loss about the use of ChatGPT. Employees are now subject to restrictions, Samsung wants to create its own alternative in the long term.

Samsung also has to learn to live and work with AI. The South Korean newspaper Economist writes that this is not going well for the technology giant for the time being. Chip division employees submitted classified information using popular AI tool ChatGPT. To prevent this, Samsung wants to develop its own AI assistant for internal use.

The Asian news outlet reports that in just twenty days there have been at least three incidents of an employee entering sensitive information into ChatGPT. This included source codes of patented applications, data from test sequences or summaries of internal meetings. Samsung has now set a limit of up to 1024 bytes per question.

reveal secrets

Sharing internal data with ChatGPT poses risks for companies. Finally, ChatGPT learns from the inputs you type into the system. For example, the tool can reproduce your codes to solve another user’s development problems. As a result, sensitive data can get into the hands of a competing company for free and at no cost.

Samsung isn’t the first company to experience this; Amazon also had to reprimand employees for disclosing secret codes. It shows that companies are looking for ways to get their employees working with AI. Since the launch of ChatGPT, the artificial intelligence debate has turned 180 degrees. Scientists are calling for a stop to development and politicians are also keeping an eye on things.