Samsung has imposed a “temporary” ban on generative AI tools like ChatGPT following what appears to be an accidental source code leak.
An internal memo seen by Bloomberg tells employees they shouldn’t use technologies like OpenAI’s ChatGPT and Google’s Bard for the pain of layoffs because of the risks to the company’s intellectual property. . The newswire said this morning that the leaked memo was sent to one of Sammy’s “largest divisions,” adding that the company has confirmed this.
“Interest in generative AI platforms such as ChatGPT is growing both domestically and internationally. While this interest focuses on the usefulness and efficiency of these platforms, there are also growing concerns about the security risks posed by generative AI. I have.”
The move appears to be a re-ban by the South Korean electronics giant. Unconfirmed reports from South Korean media that Samsung staff entered trade secrets into a chatbot while trying to fix a bug in the “problematic” source code and generate meeting minutes. A previous ban was lifted earlier this year, according to reports. suspicion of malpractice. The leaked data reportedly included equipment measurements and yield data from the conglomerate’s Device Solutions and Semiconductor business units. We asked Samsung at the time to confirm or deny this and asked again.
If we did, we wouldn’t be the first engineers to use chatbots to help with the arduous process of testing software and fixing code. Compsci researchers Chunqiu Steven Xia and Lingming Zhang have created an automated process for this, and at least he published a paper showing how one solution can be both cheap and effective. Their process, dubbed ChatRepair, not only tests patches, but also learns from previous failures and “Keeps the conversation going: fixed 162 of 337 bugs using ChatGPT for $0.42 each.” You will learn in the paper titled Do.
When it comes to OpenAI, as chatbots get more and more attention from regulators, companies developing text-based assistants have already put in place some of their own guardrails. Last week, we launched a feature that allows users to stop ChatGPT from slurping text generated in private conversations and using it to train large scale language models.
Meanwhile, under the updated policy that started on March 1st, OpenAI made two changes to its data use and retention policy.
In the EU, proposed new AI legislation could require developers to disclose and detail copyrighted data used to train ML models.
IBM, on the other hand, doesn’t seem to have any qualms about using the technology. Yesterday, CEO Arvind Krishna said that AI could replace up to 30% of IBM’s back-office operations (about 7,800 people).
Samsung is said to be working on its own AI tools. ®
