Samsung introduced ChatGPT, and within 20 days, it was reported that company secrets were leaked

Although AI artificial intelligence is smart and can help solve many troubles, it also brings many troubles. According to foreign reports, it has only been 20 days since Samsung introduced the ChatGPT chat robot, and there have been many leaks and abuses of confidential information, which may also leak company secrets to OpenAI.

In fact, in ChatGPT’s user guide, it is clearly mentioned that “users should not enter sensitive content”. This is because after the content we input in ChatGPT, the data will be sent to an external server, and these data may become OpenAI’s GPT training tool, and it is easy to cause data leakage.

However, after ChatGPT became popular, Samsung, which originally banned ChatGPT, considered that semiconductors are also the basis of generative AI artificial intelligence. After some discussions, it opened some departments to use ChatGPT on March 11, 2023.

I just didn’t expect that there would be abuse incidents within 20 days after importing. One of them was Samsung DS Semiconductor and Device Solution Department. The internal developers copied and pasted the problematic source code to ChatGPT to find out the problem and solve it. In this way, it is possible to make the source code a ChatGPT learning material.

Another employee of the DS department entered the code related to the equipment yield rate into ChatGPT, and asked ChatGPT to help optimize it, and some employees asked ChatGPT to help write meeting minutes.

According to foreign reports, when Samsung opened ChatGPT, it notified employees how to use it, and reminded that data input may be leaked, but it was only 20 days before the problem came out.

Related Articles:  Properties. The signs that anticipate the bursting of a housing bubble in the United States

Now, in order not to repeat the same mistakes, Samsung Samsung is formulating relevant measures and considering the development of Samsung Samsung’s exclusive AI; during this period, if the misuse or abuse of ChatGPT leads to possible leakage of secrets, it is not ruled out that ChatGPT will be disabled.

Original URL:three hee action wow

Image and data source: semiwiki

▶▶▶See more Sanxi action wow articles
Italy ordered to temporarily block ChatGPT, mainly due to personal data leakage
is it you? OpenAI’s research on “jobs that are more dependent on computers” is carefully replaced by AI

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.