Samsung threatens to fire employees if they leak data to A.I. chatbots like ChatGPT

Samsung said employees misusing ChatGPT risk “termination,” according to an internal memo.
Jung Yeon-Je—AFP/Getty Images

Samsung Electronics is banning employees from using programs like ChatGPT in their work, just over a month after an engineer reportedly uploaded sensitive information to the A.I. chatbot.

“There are growing concerns about security risks presented by generative A.I.,” Samsung told staff in an internal memo on Monday, reports Bloomberg.

The company cited instances of an employee leaking data to an A.I. chatbot in its memo, without giving specifics. It continued that a violation of its security policies could lead to “disciplinary action up to and including termination of employment.”

In a later announcement, Samsung clarified that the ban would be temporary. “We understand that advances in technology, such as Generative A.I. platforms, can impact future innovations that influence decisions related to product, marketing and promotional programs,” the company wrote in its revised memo, which it shared with Fortune.

In early April, Korean media reported that Samsung was restricting use of ChatGPT after several employees uploaded sensitive data to the service.

One employee reportedly asked ChatGPT to help fix a faulty database, uploading confidential source code in the process, while another shared an entire meeting in order to ask the chatbot to quickly create meeting minutes.

A.I. developers can use prompts and other submitted information to train their models.

OpenAI, for example, says it can use submissions to services like ChatGPT or image-generating model DALL-E for training. OpenAI thus asks users not to “share any sensitive information” when using ChatGPT.

Other major companies have restricted or banned the use of chatbots like ChatGPT in recent months.

As early as January, firms like Amazon were warning employees not to upload confidential information to such services. 

Even the Italian government briefly banned ChatGPT from the country, citing concerns about how the program treats personal data. (Rome has since restored access to the chatbot.)

Last week, OpenAI said it would introduce an “incognito mode” for ChatGPT, where users could opt out of having their prompts and other submissions used to train the developer’s A.I. models.

The developer also said that business subscribers would automatically opt out of having their conversations used for training purposes. 

Samsung accepted that generative A.I. programs could be a significant time-saver for employees, pledging in its memo to create its own tools for translation and document summary, reports Bloomberg

The electronics company also reportedly considered switching the default search engine used on its smartphones from Google to Microsoft’s Bing, following the latter’s embrace of generative A.I.

In the most recent quarter, Samsung’s profit dropped by 95% year on year, as the electronics company battles a slump in both semiconductor and consumer electronics spending.

The company’s chip division—normally the company’s largest profit driver—swung to a $3.4 billion loss, compared with a $6.3 billion profit a year earlier. The company forecasts demand to recover by the second half of the year.

Update, May 4, 2023: This article has been updated to note Samsung’s revised memo on its policy towards chatbots. It also clarifies that disciplinary action would be taken against employees that violate the company’s security policy.

Subscribe to Well Adjusted, our newsletter full of simple strategies to work smarter and live better, from the Fortune Well team. Sign up today.

Read More

Artificial IntelligenceCryptocurrencyMetaverseCybersecurityTech Forward