The measure also limits access to Copilot, a program from software developer GitHub that automates the composition of computer code.
Manzana has banned the use of ChatGPT among some of its employees. The US technology giant has opted for this restriction, which also affects other artificial intelligence (AI) applications, for fear of leaking confidential data, according to documents seen by ‘The Wall Street Journal’.
The measure also limits access to Copilot, a program from software developer GitHub that automates the composition of computer code. Both restrictions directly affect Microsoftwhich owns both GitHub and part of Open AI, the company that created the famous ChatGPT. Apple is developing the same technology.
The Californian multinational fears that the confidential information that users share with these systems could be leaked unintentionally. Language models like the one supported by ChatGPT are trained to interact with human users, thus responding to their queries to create all kinds of texts. To enhance each ‘conversation’, the program ‘learns’ of everything that users write. That means that, if you ask about confidential material, you can then reproduce it to other users.
Other restrictions
In March, OpenAI revealed that it had temporarily disabled ChatGPT after detecting a bug that allowed some users to see the chat history from other users. That, and its ban in Italy, led the company to allow users to have the chatbot not ‘remember’ those conversations. However, that doesn’t mean that everything we tell it will go away, OpenAI will still store some of that data.
Apple is not the only big company that has taken that step. The Korean Samsung was the first to restrict the use of generative AI calls among its employees, a path that the telephone company has also followed VerizonJP Morgan bank chase or the construction company Sacyr in Spain.