Samsung has once again “temporarily” banned the use of ChatGPT. The Korean tech giant had initially barred employees from using the AI chatbot during work, but gave it the green light again earlier this year. Not long after, Samsung staff provided sensitive company data to ChatGPT on several occasions, making the reversal a not entirely unexpected development.
Bloomberg reports that the company let personnel know of the new guidelines by way of memo. The explicit reason for the decision, then, is the risk of feeding intellectual property to generative AI models. In many cases, such as Google Bard and ChatGPT, there is a chance that the data could later be used to further train the AI.
Risk
Samsung states in the memo that it is not simply making a top-down decision: 65 percent of employees are said to agree that deploying generative AI leads to security risks. The fears expressed by staff are echoed by EU policymakers, for example: copyright and privacy are persistently challenging subjects due to the use of large data sets that generative AI needs to perform.
In the memo, Samsung asks staff to diligently adhere to the guidelines. If someone does use ChatGPT on a private basis, they may not share Samsung data in the process. The Korean company itself is developing AI tools internally, removing the IP concerns as a stumbling block for the daily use of generative AI.
Also read: Samsung employees leak secrets to ChatGPT