Itay is the first country in the West to block the advanced chatbot ChatGPT.
The Italian agency in charge of protecting personal information said there were privacy concerns with the model, which was made by the US startup OpenAI and is backed by Microsoft.
The government agency said it would ban OpenAI and look into it “immediately.”
OpenAI told the BBC that it followed laws about privacy.
Since it came out in November 2022, millions of people have used ChatGPT.
It can answer questions in a way that sounds natural and human, and it can also copy other writing styles by looking at the internet as it was in 2021.
It cost Microsoft billions of dollars to make, and Bing got it last month.
It has also said that it will put a version of the technology into its Office apps, such as Word, Excel, PowerPoint, and Outlook.
Concerns have been raised about the possible dangers of artificial intelligence (AI), such as the fact that it could threaten jobs and help spread false information and bias.
This week, big names in tech, like Elon Musk, called for a halt to the development of these kinds of AI systems because they worried that the race to make them was getting out of hand.
The Italian watchdog said that it would stop OpenAI’s chatbot and check to see if it followed the General Data Protection Regulation.
GDPR tells us how we can use, process, and store information about people.
On March 20, the watchdog said that the app had a data breach that exposed users’ conversations and payment information.
It said that there was no legal justification for “the mass collection and storage of personal data to “train” the algorithms that run the platform.”
It also said that there was no way to check a user’s age, so the app “exposes minors to answers that are completely inappropriate for their level of development and awareness.”
Bard, Google’s rival chatbot with artificial intelligence, is now available, but it can only be used by people over the age of 18.
The Italian data protection authority said OpenAI had 20 days to explain how it would address the watchdog’s concerns or face a fine of up to €20 million ($21.7 million) or 4% of its annual revenue.
Also, the Irish data protection commission told the BBC that it is following up with the Italian regulator to find out why it did what it did and that it “will coordinate with all EU data protection authorities” about the ban.
The UK’s independent data regulator, the Information Commissioner’s Office, told the BBC that it would “support” developments in AI but would also “challenge” companies that don’t follow data protection laws.
Dan Morgan from the cybersecurity ratings company SecurityScorecard said that the ban shows how important it is for companies doing business in Europe to follow the rules.
“Businesses must put the protection of personal data at the top of their list of priorities and follow the strict data protection rules set by the EU. Following the rules is not an option.”
“Not well enough regulated”
After a complaint was filed in the US, the consumer advocacy group BEUC also asked EU and national authorities, such as data protection watchdogs, to look into ChatGPT and other similar chatbots.
Even though the EU is working on the world’s first AI law, BEUC is worried that it will take years for the AI Act to go into effect, leaving consumers at risk of harm from a technology that isn’t regulated well enough.
Ursula Pachl, who is the deputy director general of BEUC, said that society is “not protected enough from the harm” AI can cause right now.
She said, “There are growing worries about how ChatGPT and other chatbots could trick and control people. These AI systems need more public scrutiny, and public authorities must get back in charge of them.”
ChatGPT is already blocked in China, Iran, North Korea, and Russia, among other places.
OpenAI told the BBC that it had turned off ChatGPT for people in Italy at the request of the Garante, which is Italy’s data protection regulator.
It said, “We are committed to protecting people’s privacy, and we think we follow GDPR and other privacy laws.”
The organisation said it worked to reduce the amount of personal data used to train AI systems like ChatGPT because it wanted its AI systems to “learn about the world, not about private individuals.”
“We also believe that AI regulation is necessary, so we look forward to working closely with the Garante and teaching them how our systems are built and used,” it said.
OpenAI was excited to “soon” make ChatGPT available again in Italy.