Privacy

ChatGPT is violating Europe’s privacy laws, Italian DPA tells OpenAI

Comment

OpenAI logo is being displayed on a mobile phone screen in front of computer screen with the logo of ChatGPT
Image Credits: Didem Mente/Anadolu Agency / Getty Images

OpenAI has been told it’s suspected of violating European Union privacy, following a multi-month investigation of its AI chatbot, ChatGPT, by Italy’s data protection authority.

Details of the Italian authority’s draft findings haven’t been disclosed. But the Garante said today OpenAI has been notification and given 30 days to respond with a defence against the allegations.

Confirmed breaches of the pan-EU regime can attract fines of up to €20 million, or up to 4% of global annual turnover. More uncomfortably for an AI giant like OpenAI, data protection authorities (DPAs) can issue orders that require changes to how data is processed in order to bring an end to confirmed violations. So it could be forced to change how it operates. Or pull its service out of EU Member States where privacy authorities seek to impose changes it doesn’t like.

OpenAI was contacted for a response to the Garante’s notification of violation. We’ll update this report if they send a statement.

Update: OpenAI said:

We believe our practices align with GDPR and other privacy laws, and we take additional steps to protect people’s data and privacy. We want our AI to learn about the world, not about private individuals. We actively work to reduce personal data in training our systems like ChatGPT, which also rejects requests for private or sensitive information about people. We plan to continue to work constructively with the Garante.

AI model training lawfulness in the frame

The Italian authority raised concerns about OpenAI’s compliance with the bloc’s General Data Protection Regulation (GDPR) last year — when it ordered a temporary ban on ChatGPT’s local data processing which led to the AI chatbot being temporarily suspended in the market.

The Garante’s March 30 provision to OpenAI, aka a “register of measures”, highlighted both the lack of a suitable legal basis for the collection and processing of personal data for the purpose of training the algorithms underlying ChatGPT; and the tendency of the AI tool to ‘hallucinate'(i.e. its potential to produce inaccurate information about individuals) — as among its issues of concern at that point. It also flagged child safety as a problem.

In all, the authority said that it suspected ChatGPT to be breaching Articles 5, 6, 8, 13 and 25 of the GDPR.

Despite identifying this laundry list of suspected violations, OpenAI was able to resume service of ChatGPT in Italy relatively quickly last year, after taking steps to address some issues raised by the DPA. However the Italian authority said it would continue to investigate the suspected violations. It’s now arrived at preliminary conclusions the tool is breaking EU law.

While the Italian authority hasn’t yet said which of the previously suspected ChatGPT breaches it’s confirmed at this stage, the legal basis OpenAI claims for processing personal data to train its AI models looks like a particularly crux issue.

This is because ChatGPT was developed using masses of data scraped off the public Internet — information which includes the personal data of individuals. And the problem OpenAI faces in the European Union is that processing EU people’s data requires it to have a valid legal basis.

The GDPR lists six possible legal bases — most of which are just not relevant in its context. Last April, OpenAI was told by the Garante to remove references to “performance of a contract” for ChatGPT model training — leaving it with just two possibilities: Consent or legitimate interests.

Given the AI giant has never sought to obtain the consent of the countless millions (or even billions) of web users’ whose information it has ingested and processed for AI model building, any attempt to claim it had Europeans’ permission for the processing would seem doomed to fail. And when OpenAI revised its documentation after the Garante’s intervention last year it appeared to be seeking to rely on a claim of legitimate interest. However this legal basis still requires a data processor to allow data subjects to raise an objection — and have processing of their info stop.

How OpenAI could do this in the context of its AI chatbot is an open question. (It might, in theory, require it to withdraw and destroy illegally trained models and retrain new models without the objecting individual’s data in the training pool — but, assuming it could even identify all the unlawfully processed data on a per individual basis, it would need to do that for the data of each and every objecting EU person who told it to stop… Which, er, sounds expensive.)

Beyond that thorny issue, there is the wider question of whether the Garante will finally conclude legitimate interests is even a valid legal basis in this context.

Frankly, that looks unlikely. Because LI is not a free-for-all. It requires data processors to balance their own interests against the rights and freedoms of individuals whose data is being processed — and to consider things like whether individuals would have expected this use of their data; and the potential for it to cause them unjustified harm. (If they would not have expected it and there are risks of such harm LI will not be found to be a valid legal basis.)

The processing must also be necessary, with no other, less intrusive way for the data processor to achieve their end.

Notably, the EU’s top court has previously found legitimate interests to be an inappropriate basis for Meta to carry out tracking and profiling of individuals to run its behavioral advertising business on its social networks. So there is a big question mark over the notion of another type of AI giant seeking to justify processing people’s data at vast scale to build a commercial generative AI business — especially when the tools in question generate all sorts of novel risks for named individuals (from disinformation and defamation to identity theft and fraud, to name a few).

A spokesperson for the Garante confirmed that the legal basis for processing people’s data for model training remains in the mix of what it’s suspected ChatGPT of violating. But they did not confirm exactly which one (or more) article(s) it suspects OpenAI of breaching at this point.

The authority’s announcement today is also not yet the final word — as it will also wait to receive OpenAI’s response before taking a final decision.

Here’s the Garante’s statement (which we’ve translated from Italian using AI):

[Italian Data Protection Authority] has notified OpenAI, the company that runs the ChatGPT artificial intelligence platform, of its notice of objection for violating data protection regulations.

Following the provisional restriction of processing order, adopted by the Garante against the company on March 30, and at the outcome of the preliminary investigation carried out, the Authority considered that the elements acquired may constitute one or more unlawful acts with respect to the provisions of the EU Regulation.

OpenAI, will have 30 days to communicate its defence briefs on the alleged violations.

In defining the proceedings, the Garante will take into account the ongoing work of the special task force set up by the Board that brings together the EU Data Protection Authorities (EDPB).

OpenAI is also facing scrutiny over ChatGPT’s GDPR compliance in Poland, following a complaint last summer which focuses on an instance of the tool producing inaccurate information about a person and OpenAI’s response to that complainant. That separate GDPR probe remains ongoing.

OpenAI, meanwhile, has responded to rising regulatory risk across the EU by seeking to establish a physical base in Ireland; and announcing, in January, that this Irish entity would be the service provider for EU users’ data going forward.

Its hopes with these moves will be to gain so-called “main establishment” status in Ireland and switch to having assessment of its GDPR compliance led by Ireland’s Data Protection Commission, via the regulation’s one-stop-shop mechanism — rather than (as now) its business being potentially subject to DPA oversight from anywhere in the Union that its tools have local users.

However OpenAI has yet to obtain this status so ChatGPT could still face other probes by DPAs elsewhere in the EU. And, even if it gets the status, the Italian probe and enforcement will continue as the data processing in question predates the change to its processing structure.

The bloc’s data protection authorities have sought to coordinate on their oversight of ChatGPT by setting up a taskforce to consider how the GDPR applies to the chatbot, via the European Data Protection Board, as the Garante’s statement notes. That (ongoing) effort may, ultimately, produce more harmonized outcomes across discrete ChatGPT GDPR investigations — such as those in Italy and Poland.

However authorities remain independent and competent to issue decisions in their own markets. So, equally, there are no guarantees any of the current ChatGPT probes will arrive at the same conclusions.

ChatGPT resumes service in Italy after adding privacy disclosures and controls

Italy gives OpenAI initial to-do list for lifting ChatGPT suspension order

 

More TechCrunch

WhatsApp is updating its mobile apps for a fresh and more streamlined look, while also introducing a new “darker dark mode,” the company announced on Thursday. The messaging app says…

WhatsApp’s latest update streamlines navigation and adds a ‘darker dark mode’

Plinky lets you solve the problem of saving and organizing links from anywhere with a focus on simplicity and customization.

Plinky is an app for you to collect and organize links easily

The keynote kicks off at 10 a.m. PT on Tuesday and will offer glimpses into the latest versions of Android, Wear OS and Android TV.

Google I/O 2024: How to watch

For cancer patients, medicines administered in clinical trials can help save or extend lives. But despite thousands of trials in the United States each year, only 3% to 5% of…

Triomics raises $15M Series A to automate cancer clinical trials matching

Welcome back to TechCrunch Mobility — your central hub for news and insights on the future of transportation. Sign up here for free — just click TechCrunch Mobility! Tap, tap.…

Tesla drives Luminar lidar sales and Motional pauses robotaxi plans

The newly announced “Public Content Policy” will now join Reddit’s existing privacy policy and content policy to guide how Reddit’s data is being accessed and used by commercial entities and…

Reddit locks down its public data in new content policy, says use now requires a contract

Eva Ho plans to step away from her position as general partner at Fika Ventures, the Los Angeles-based seed firm she co-founded in 2016. Fika told LPs of Ho’s intention…

Fika Ventures co-founder Eva Ho will step back from the firm after its current fund is deployed

In a post on Werner Vogels’ personal blog, he details Distill, an open-source app he built to transcribe and summarize conference calls.

Amazon’s CTO built a meeting-summarizing app for some reason

Paris-based Mistral AI, a startup working on open source large language models — the building block for generative AI services — has been raising money at a $6 billion valuation,…

Sources: Mistral AI raising at a $6B valuation, SoftBank ‘not in’ but DST is

You can expect plenty of AI, but probably not a lot of hardware.

Google I/O 2024: What to expect

Dating apps and other social friend-finders are being put on notice: Dating app giant Bumble is looking to make more acquisitions.

Bumble says it’s looking to M&A to drive growth

When Class founder Michael Chasen was in college, he and a buddy came up with the idea for Blackboard, an online classroom organizational tool. His original company was acquired for…

Blackboard founder transforms Zoom add-on designed for teachers into business tool

Groww, an Indian investment app, has become one of the first startups from the country to shift its domicile back home.

Groww joins the first wave of Indian startups moving domiciles back home from US

Technology giant Dell notified customers on Thursday that it experienced a data breach involving customers’ names and physical addresses. In an email seen by TechCrunch and shared by several people…

Dell discloses data breach of customers’ physical addresses

Featured Article

Fairgen ‘boosts’ survey results using synthetic data and AI-generated responses

The Israeli startup has raised $5.5M for its platform that uses “statistical AI” to generate synthetic data that it says is as good as the real thing.

4 hours ago
Fairgen ‘boosts’ survey results using synthetic data and AI-generated responses

Hydrow, the at-home rowing machine maker, announced Thursday that it has acquired a majority stake in Speede Fitness, the company behind the AI-enabled strength training machine. The rowing startup also…

Rowing startup Hydrow acquires a majority stake in Speede Fitness as their CEO steps down

Call centers are embracing automation. There’s debate as to whether that’s a good thing, but it’s happening — and quite possibly accelerating. According to research firm TechSci Research, the global…

Retell AI lets companies build ‘voice agents’ to answer phone calls

TikTok is starting to automatically label AI-generated content that was made on other platforms, the company announced on Thursday. With this change, if a creator posts content on TikTok that…

TikTok will automatically label AI-generated content created on platforms like DALL·E 3

India’s mobile payments regulator is likely to extend the deadline for imposing market share caps on the popular UPI (unified payments interface) payments rail by one to two years, sources…

India likely to delay UPI market caps in win for PhonePe-Google Pay duopoly

Line Man Wongnai, an on-demand food delivery service in Thailand, is considering an initial public offering on a Thai exchange or the U.S. in 2025.

Thai food delivery app Line Man Wongnai weighs IPO in Thailand, US in 2025

The problem is not the media, but the message.

Apple’s ‘Crush’ ad is disgusting

Ever wonder why conversational AI like ChatGPT says “Sorry, I can’t do that” or some other polite refusal? OpenAI is offering a limited look at the reasoning behind its own…

OpenAI offers a peek behind the curtain of its AI’s secret instructions

The federal government agency responsible for granting patents and trademarks is alerting thousands of filers whose private addresses were exposed following a second data spill in as many years. The…

US Patent and Trademark Office confirms another leak of filers’ address data

As part of an investigation into people involved in the pro-independence movement in Catalonia, the Spanish police obtained information from the encrypted services Wire and Proton, which helped the authorities…

Encrypted services Apple, Proton and Wire helped Spanish police identify activist

Match Group, the company that owns several dating apps, including Tinder and Hinge, released its first-quarter earnings report on Tuesday, which shows that Tinder’s paying user base has decreased for…

Match looks to Hinge as Tinder fails

Private social networking is making a comeback. Gratitude Plus, a startup that aims to shift social media in a more positive direction, is expanding its wellness-focused, personal reflections journal to…

Gratitude Plus makes social networking positive, private and personal

With venture totals slipping year-over-year in key markets like the United States, and concern that venture firms themselves are struggling to raise more capital, founders might be worried. After all,…

Can AI help founders fundraise more quickly and easily?

Google has found a way to bring a variation of its clever “Circle to Search” gesture to iPhone users. The new interaction, launched in January, allows Android users to search…

Google brings a variation on ‘Circle to Search’ to iPhone users

A new sculpture going live on Wednesday in the Flatiron South Public Plaza in New York is not your typical artwork. It combines technology, sociology, anthropology and art to let…

Always-on video portal lets people in NYC and Dublin interact in real time

Apple’s iPad event had a lot to like. New iPads with new chips and new sizes, a new Apple Pencil, and even some software updates. If you are a big…

TechCrunch Minute: When did iPads get as expensive as MacBooks?