Regulators announced Thursday that the company behind ChatGPT would submit ways to address the data protection issues that prompted a temporary Italian ban on the artificial intelligence chatbot. This means that OpenAI to profer solutions to Italy’s ChatGPT ban.
Last week, the famous San Francisco-based OpenAI chatbot was disabled by the Italian data protection agency, known as Garante, which also ordered it to temporarily halt processing the personal data of Italian users while it looks into a potential violation of EU data privacy laws.
According to experts, it was the first instance of a democracy placing a broad national ban on a popular AI platform. The business vowed to outline steps to remedy the issues in a video chat late on Wednesday between the watchdog’s commissioners and OpenAI leaders, including CEO Sam Altman. These cures have yet to be described in detail.
The Italian watchdog claimed it didn’t want to obstruct AI’s advancement but emphasized to OpenAI the significance of adhering to the strict privacy regulations set forth by the EU’s 27 member nations.
After some users’ messages and payment information were made public, the regulators issued the suspension. They also questioned whether OpenAI had the right to gather the vast amounts of data necessary to train ChatGPT’s algorithms and expressed concern that the system would occasionally produce inaccurate data on certain people.
The so-called generative AI technology, like ChatGPT, can produce prose closely resembling human writing styles since it has been “trained” on vast data sets, including digital books and online writings.
These technologies have generated excitement in the tech community and beyond, but they have also raised concerns about potential ethical and societal implications among authorities, regulators, computer scientists, and tech sector leaders.
After Italy’s move, other regulators in Europe and internationally have begun to pay closer attention. The Irish Data Protection Commission stated that it is “keeping up” with the Italian regulator to learn the reasons behind their response and will “coordinate with other EU Data Protection Authorities in connection to this subject.”
After receiving two complaints regarding ChatGPT, France’s data privacy watchdog, CNIL, announced that it is conducting an investigation. After a complaint on the alleged “collection, use, and disclosure of personal information without consent,” Canada’s privacy commissioner has also launched an investigation into OpenAI.
In a blog post, the U.K. Information Commissioner’s Office warned that “organizations developing or employing generative AI should be considering their data protection duties from the outset” and build systems with data protection as a default.
The office stated, “This isn’t voluntary – if you’re processing personal data, it’s the law.” OpenAI issued a blog post on Wednesday describing its approach to AI safety, seemingly in reaction to the worries. The company claimed that it works to remove personally identifiable information from training data when practical.
It improves its models to reject requests for personally identifiable information of private individuals and responds to requests to delete personally identifiable information from its systems.
Italy Temporarily Bans ChatGPT On Privacy Concerns
Italy’s ChatGPT ban, the government’s privacy watchdog announced Friday that ChatGPT, an artificial intelligence software, has been temporarily blocked after a data breach. The Italian Data Protection Authority temporarily barred ChatGPT from processing Italian users’ data “until ChatGPT respects privacy.”
U.S.-based OpenAI blocked ChatGPT for Italian users late Friday night at the government’s request. The business complies with European privacy rules and wants to return ChatGPT soon.
Italy’s ban on ChatGPT is “the first nation-scale limitation of a mainstream AI platform by a democracy,” according to NetBlocks director Alp Toker.
The limitation affects the ChatGPT online application, popular writing assistance, but not software apps from companies like Microsoft’s Bing search engine that have licenses with OpenAI to use the chatbot’s technology.
Large language models, which power chatbots, can emulate human writing styles because they have consumed much internet writing and digital books. The Italian authority ordered OpenAI to report its data protection procedures within 20 days or risk a fine of 20 million euros (almost $22 million) or 4% of worldwide income.
The agency cited the EU’s General Data Protection Regulation and a ChatGPT data breach involving “users’ interactions” and subscriber payments. OpenAI previously announced that ChatGPT would be unavailable on March 20 to repair an issue that allowed certain users to access other users’ chat history titles.
The company reported that 1.2% of ChatGPT Plus users may have shared personal data. “The number of users whose data was leaked to someone else is extremely minimal and we reached out to everybody who might be affected.”
The Garante, Italy’s privacy authority, questioned if OpenAI’s “bulk gathering and processing of personal data” to train its algorithms was allowed. ChatGPT can also produce and store incorrect information about people.
Finally, there’s no age verification system, exposing children to “totally inappropriate to their age and awareness” comments. “We want our AI to learn about the globe, not private individuals,” OpenAI replied.
“AI regulation is vital — so we look forward to working closely with the Garante and teaching them on how our systems are designed and used,” the business added on Italy’s ChatGPT ban.
Concerns about the AI rise prompted the Italian watchdog’s approach. A group of scientists and tech industry experts signed a statement Wednesday urging businesses like OpenAI to suspend the development of more powerful AI models until fall to allow society to assess the hazards.
Friday evening, the president of Italy’s privacy watchdog agency told state TV he signed the appeal. Pasquale Stanzione did so because “it’s not apparent what purposes are being pursued” by AI developers. Stanzione added, “this is serious” if AI “impinges” on self-determination. He called the lack of filters for under-13s “very grave.”
Italy’s ChatGPT ban, last week, San Francisco-based OpenAI CEO Sam Altman announced a six-continent journey to discuss the technology with users and developers in May. That includes stops in Brussels, Madrid, Munich, London, and Paris, where EU lawmakers have debated sweeping new laws to curb high-risk AI products.
BEUC, a European consumer group, urged the EU and its 27 member states to investigate ChatGPT and other AI chatbots on Thursday. BEUC warned that EU AI laws might not take effect for years. Therefore, authorities must act quickly to protect consumers.
“In a few months, we have witnessed a significant take-up of ChatGPT, and this is only the beginning,” stated Deputy Director General Ursula Pachl. “Waiting for the EU’s AI Act isn’t good enough since there are genuine concerns arising about how ChatGPT and similar chatbots might deceive and manipulate people.”
Conclusion
Regulators announced on Thursday that ChatGPT’s developer would submit solutions to the data protection issues that led Italy to prohibit the artificial intelligence chatbot temporarily. The popular San Francisco-based OpenAI chatbot was barred last week after being ordered to temporarily halt processing the personal data of Italian users while Garante, the Italian data protection office, looks into a potential violation of EU data privacy laws. According to experts, it was the first instance of a democracy placing a broad ban on a popular AI platform.
On Italy’s ChatGPT ban, OpenAI executives, including CEO Sam Altman, agreed to outline steps to allay the worries in a late-night video chat with the watchdog’s commissioners. These treatments have not been specified. Although the Italian watchdog claimed it did not wish to impede AI development, it emphasized to OpenAI the necessity of adhering to the strict privacy regulations set down by the EU’s 27 member nations. Following the exposure of certain users’ messages and payment information to outsiders, the regulators issued the ban.
