Italy will say ciao to ChatGPT ban if OpenAI does indeed think of the children
And also, privacy safeguards
Updated Italy's privacy watchdog will lift the country's ban on ChatGPT if OpenAI agrees to enforce rules protecting minors and users' personal data by April 30, it announced on Wednesday.
The Guarantor for the Protection of Personal Data (GPDP) announced it had blocked access to ChatGPT to impose a "temporary limitation of the processing of personal data" last month. The ban was put in place as regulators began investigating whether the chatbot violated the European Union's GDPR and Italy's data privacy laws.
Officials were concerned the large language model, trained on huge amounts of text scraped from the internet, could recall and emit personal information, such as people's phone numbers or addresses, from input queries. Users may be disclosing sensitive information about themselves or others to the chatbot too, and unexpectedly allowing OpenAI to collect and store that private data.
Like all AI text-generating models, ChatGPT can produce toxic, biased, and false information. The GPDP was also afraid the chatbot could expose minors to inappropriate answers that could potentially be harmful given their "degree of development and self-awareness."
Now, regulators have listed a number of solutions to address their data privacy concerns, and will unblock ChatGPT if OpenAI agrees to implement the corresponding safeguards in its software.
"OpenAI will have to comply by April 30 with the measures set out by the Italian [Supervisory Authority (SA)] concerning transparency, the right of data subjects – including users and non-users - and the legal basis of the processing for algorithmic training relying on users' data," the GPDP said in a statement.
"Only in that case will the Italian SA lift its order that placed a temporary limitation on the processing of Italian users' data, there being no longer the urgency underpinning the order, so that ChatGPT will be available once again from Italy."
- Canada sticks a privacy probe into OpenAI's ChatGPT
- Euro privacy regulators sniff Italy's ChatGPT ban, consider a pizza the action
- Italy bans ChatGPT for 'unlawful collection of personal data'
- AI conference and NYC's educators ban papers done by ChatGPT
OpenAI has been asked to notify people on its website how ChatGPT stores and processes their data, and require users to confirm they're 18 and older before using the software. The biz must also ask for explicit consent to use people's data to train their AI models, and allow anyone, user or not, to request any false personal information generated by ChatGPT to be corrected or deleted altogether.
Won't kids just lie about their age? Well, OpenAI must also roll out an age verification process when registering new users, and also prevent children below the age of 13 from accessing the software. People aged 13 to 18 must obtain consent to use ChatGPT from their parents, and all these changes must be implemented by September 30, or else the ban stays.
This latter policy is largely in line with OpenAI's previously stated plans, but setting up a fact-checking and deletion system could be a major headache if the April 30 deadline is to be met.
Finally, if Open AI plays ball with the proposed deadlines, it must run a nationwide advertising campaign on "radio, television, newspapers and the web" to let everyone know the good news.
Italy's not the only country to investigate privacy shortcomings in ChatGPT. Regulators in Canada and France announced it was probing whether the software was unlawfully collecting, storing, and using personal data after receiving official complaints. Meanwhile Spain has urged the EU's privacy watchdog to examine ChatGPT further, according to Reuters.
The Register has asked OpenAI for comment. ®
Updated to add
"We are happy that the Italian Garante is reconsidering their decision and we look forward to working with them to make ChatGPT available to our customers in Italy again soon," OpenAI told The Register.