ChatGPT Waterloo: Widespread account ban; banned by Italy

ChatGPT seems to be encountering a Waterloo.

On March 31, a large number of users complained on social platforms that they could no longer log in to the ChatGPT account they had opened with their own money, and what was even more speechless was that they could not register a new account. As more and more users revealed, it was discovered that the account ban incident had started on March 30 and was spreading rapidly.

On April 1, more and more people said that their accounts had been blocked, and people on major platforms and group chats rushed to tell each other "notes" to prevent being blocked. At this time, the Plus accounts that were originally usable were not spared, and became invalid one after another, and some records could not be checked in the bill.

Strangely, as of April 3, the large-scale account bans did not seem to have stopped. Many netizens also posted on social platforms saying, "ChatGPT is undergoing a large-scale ban." The area is closed." A large number of users reported that the silent account bans this time were mainly concentrated in Asia. At the same time, OpenAI did not issue any warning or response to this.

In terms of compliance, ChatGPT has also encountered new scrutiny. The Record website disclosed that the Italian data protection agency officially announced a temporary ban on the use of ChatGPT and strictly requested its parent company OpenAI to stop processing Italian user information. Italy "fired" the first shot to ban ChatGPT!

As for the reasons for banning ChatGPT, the Italian data protection agency stated that OpenAI not only illegally collected a large amount of personal data of Italian users, but what is even worse is that OpenAI did not set up a mechanism to check the age of ChatGPT users to prevent minors from accessing it. Illegal material.

The Italian data protection agency further stated that OpenAI has 20 days to defend its theft of information, otherwise it may be fined up to 20 million euros or the company's annual turnover for violating the EU's General Data Protection Regulation 4% penalty. Not only in terms of fines, once OpenAI cannot explain in detail whether it has taken measures to protect user data information, it will inevitably increase the probability of EU countries banning ChatGPT.

After Italy banned ChatGPT, OpenAI immediately took ChatGPT offline in Italy. Its CEO Sam Altman immediately tweeted to explain that the work done by the company was to train artificial intelligence systems such as ChatGPT. When reducing personal data, we hope that our artificial intelligence understands the world rather than individuals. The company has not violated any regulations and very much looks forward to cooperating with the Italian personal data protection authority.

ChatGPT is undoubtedly a double-edged sword

It is undeniable that the emergence of ChatGPT will set off a new round of information revolution and better benefit mankind, but any Everything has two sides. When society looks forward to the unprecedented opportunities that ChatGPT will bring to mankind, it should also think about the security risks that its emergence will bring. After all, in just two or three months, security risks arising from the use of ChatGPT have emerged one after another.

Users can view the chat records of other users

ChatGPT has always presented itself as a "perfect" image since its launch, but as the number of users gradually increased, it gradually began to expose some security issues Hidden danger.

In early March 2023, some users revealed that other people’s chat records were displayed in their chat lists. For a while, ChatGPT leaked user data and occupied major social media lists. In the end, its parent company OpenAI CEO Sam... Altman had to urgently issue a tweet to explain to users that the BUG that caused the user's ChatGPT chat history to be leaked was just a bug in the open source code base.

After the leak of user chat records, although OpenAI readily admitted to the media that there was a security vulnerability in ChatGPT and immediately went offline Chat history function, but it has been emphasized that only the chat history title is leaked, and users cannot see other people’s actual chat content.

Judging from the information disclosed by OpenAI, user conversation content will be automatically saved to the cloud, and there is the possibility of being stolen by a third party. Combined with this historical chat record leak incident, more and more security researchers have begun to think seriously. Whether information security issues in ChatGPT are accidental.

Samsung’s use of ChatGPT led to the leakage of confidential data

One wave after another. Last week, South Korean media revealed that technology giant Samsung Electronics was exposed to three leaks of internal confidential data less than 20 days after introducing ChatGPT. (2 cases were related to semiconductor equipment, and 1 case was related to the content of internal meetings).

It is reported that these confidential information of Samsung Electronics may have been imported into the ChatGPT database and are at risk of leakage at any time.

Details of the three incidents:

• Samsung DS Equipment Solutions Department employees discovered the existence of copies when operating semiconductor test equipment to download software There were some problems, so I pasted the problematic code into ChatGPT and tried to get help. However, this may cause ChatGPT to use Samsung’s confidential information as training materials;
• Another person in the DS Device Solutions Department An employee tried to use ChatGPT to help him optimize a piece of code, so he directly entered a piece of code related to output and yield recording equipment;
• A Samsung insider used the AI ​​voice assistant Naver Clova to record the meeting After converting it into text, I want to use ChatGPT to summarize and refine the meeting content, but this may lead to the leakage of confidential meeting minutes.

After the Samsung Electronics leak, they immediately organized security personnel internally to formulate relevant protective measures, and stated that if similar incidents occur again in the future, Samsung may cut off the ChatGPT service internally. In addition, the media revealed that Samsung is currently considering developing Samsung-specific AI services through the DS Department Innovation Center.

On March 11 this year, Samsung Electronics passed a resolution to allow DS department employees to use ChatGPT, but ChatGPT is still disabled for DX device experience department employees. As a result, we encountered three unexpected security incidents in just twenty days. This shows that the security of ChatGPT is indeed worthy of consideration.

Samsung is not the only company where employees upload corporate confidential data to ChatGPT. Judging from Cyberhaven's statistics on the use of ChatGPT by 1.6 million employees, about 3.1% of employees directly upload internal corporate data to ChatGPT for analysis in order to improve work efficiency.

Many scientific figures have jointly called for the suspension of more powerful AI development research

As can be seen from the above mentioned content, ChatGPT is widely used and brings some Serious security risks, therefore, regarding ChatGPT and the future of artificial intelligence, have aroused public concerns about AI technology, and many well-known scientific figures have jointly stated that they have suspended the development of more powerful AI systems.

On March 29, US time, including the father of artificial intelligence, Jeffrey Hinton, Tesla and Twitter CEO Elon Musk, Turing Award winner Joshua Bengio, etc. Thousands of artificial intelligence experts and industry executives, including the company, signed an open letter titled "Suspension of Large Model Experiments in AI" released by the Future of Life Institute, a US non-profit organization.

In the letter, these senior experts called on all artificial intelligence laboratories to suspend the development and training of more powerful artificial intelligence systems for at least six months. Under special circumstances, governments should Intervene promptly and immediately suspend AI system research.

The above is the detailed content of ChatGPT Waterloo: Widespread account ban; banned by Italy. For more information, please follow other related articles on the PHP Chinese website!