Scorza (Privacy Guarantor): “Against OpenAI and ChatGpt emergency measure”. Here’s what can happen now

“We have asked OpenAi to suspend the processing of personal data of Italian users with immediate effect. We acted independently. Now the company has 20 days to comply. Guido Scorza since 2020 he has been a member of the board of the authority for the protection of privacy. The company that created ChatGpt, according to the panel, would not comply with the Italian law on data processing. “There is no disclosure, and this is the offense that we contest against the company. But it is evident that there is also another theme: with Chatgpt and the chatbots we have conversations, and in these conversations we often tend to share a lot of our lives”.

Lawyer, what do you dispute with OpenAI?

“The fact that on their site there is no information on the processing of personal data of Italian users. At ChatGpt you can ask when Paolo Rossi was born, and get an answer. This is personal data. Every question about people’s lives is data processing. Not to mention those he treats ‘wrongly’”.

Can you give us an example?

“If I ask the chatbot when Guido Scorza entered the board of the Privacy Guarantor, it says it was 2016. But I entered it in 2020. In addition to the unlawful processing of data, there is also that the processing is inaccurate in many cases” .

What does your stop entail?

“In the meantime, OpenAi must stop processing the personal data present in the material fed to artificial intelligences to improve itself. We have adopted an emergency measure. Which is generally taken on the basis of a ‘fumus’, therefore not on the certainty that there is an offense but on the apparent legitimacy of the suspicion that there is. Then it will depend on how the company responds”.

Can we think of an inaccessible ChatGpt in Italy?

“Inaccessibility is not an order that we decide. If from tomorrow ChatGpt manages to work without processing personal data, it can remain accessible. But without reporting information on people’s lives, nor answering questions such as ‘where was Paolo Rossi born’”.

So what happens now?

“In the meantime, an investigation has been launched. Which will continue anyway. It must be ascertained whether there has been a violation or not. The best thing would be to do it together with the company, in this case OpenAI. But even if they don’t agree to cooperate and provide us with their reasons, the investigation will continue”.

What are the possible scenarios?

“The preliminary investigation could confirm the lawfulness or otherwise of OpenAI on data processing. In the first case there is archiving. In the second, measures could be taken: either corrective or sanctioning”.

What if your investigation confirms that OpenAI is behaving illegally?

“Should we decide on a corrective measure, OpenAI could be forced to cancel the data accumulated on Italian users, or not to collect any more in the future. As for the sanctions, in this case we are talking about economic sanctions, which in the case of Italy range from 2 to 4% of the turnover of the company to which the offense is contested”.

In addition to processing and collecting data, ChatGpt is a tool used by many to ask questions, sometimes intimate, or concerning the strictly personal sphere. Was this also what prompted you to move urgently?

“It is a synthesis of the two things. There is a strong suspicion that there is an offense. And we will clarify it with the investigation. Here without a doubt the offense could have more serious consequences because through a conversation one is led to share even more things concerning the intimate sphere. But basically there is the violation of a rule”.

Did you act autonomously or is there some kind of coordination between European privacy authorities?

“In this case we acted independently because OpenAi is not based in Europe and is outside the cooperation procedures. But in the next few days I imagine there will be room for discussion at a European level to understand if and how to cooperate and act together”.