The bloc’s omission hints at the larger power struggle between regulators and Big Tech in Europe. At a minimum, the search giant needs more time to compare Bard with the EU’s draft AI Act, says Henk van Ess, a visiting professor at the Freetech Axel Springer Academy of Journalism and Technology in Berlin. “The proposed regulation emphasizes the importance of transparency and traceability in AI systems,” he says. “It may be challenging for large language models like Google Bard to fully comply with this requirement, as the decision-making process in these models can be complex and not easily interpretable.”
AI Act rules may also pose an issue if Google has trained Bard on a data set containing errors or biases, adds van Ess. In April, researchers found that they could prompt Bard to deny climate change, mischaracterize the war in Ukraine, and question vaccine efficacy. “Google is playing cautious,” says Robin Rohm, founder of Berlin-based AI startup Apheris. “They recognize that Bard might be considered a product that enables high-risk applications, and this could expose them to risk under the proposed regulation. The delay could be an effort to buy them time.”
Any high-profile mistakes in the EU could cost the company big in the months and years to come. Google will be sensitive to the fact that whatever happens now will likely influence the negotiations around the AI Act, according to Daniel Leufer, a Brussels-based senior policy analyst with digital rights group Access Now. “If ChatGPT, Bard, et cetera, in the next six to seven months, are responsible for serious public blunders, then measures that would address those blunders could very well find their way into the AI Act,” he says.
The draft rules of the AI Act aren’t expected to be approved until next year, but other EU regulations may already be causing Google headaches. Europe’s new Digital Services Act could also be playing a role when it comes to Google incorporating Bard into its search setup, says Harshvardhan Pandit, an assistant professor at the Adapt Centre at Dublin City University. “Given that Bard also acts as a search engine, it may be that Google is also trialing integrating ads into it and does not want to be subject to the DSA at the moment,” Pandit says. The DSA introduces new rules around online advertising.
With competition to push out more generative AI services building, Europe’s privacy laws are already causing issues for new services. “There’s a lingering question whether these very large data sets, that have been collected more or less by indiscriminate scraping, have a sufficient legal basis under the GDPR,” says Leufer. At the end of March, Italy’s data regulator temporarily banned ChatGPT for not following the bloc’s GDPR privacy rules. OpenAI had “unlawfully” collected personal information from the web as part of its training data, the regulator said, as well as failing to inform people how their data was being used or developing tools to stop children from using ChatGPT.
The move resulted in OpenAI making changes to let people delete more data from ChatGPT, along with a host of other concessions. While ChatGPT is now available in Italy again, the country’s data regulator is still scrutinizing the technology. The decision prompted other EU countries to create a joint task force to investigate ChatGPT further. The Irish Data Protection Authority, which handles GDPR issues relating to Google, Meta, Microsoft and Apple, didn’t immediately respond to a request for comment about whether it had discussed the rollout of Bard in Europe with Google.
However, it’s possible that GDPR rules are one reason Bard isn’t being launched in the European Economic Area (EEA), a group of countries that includes the EU bloc, Norway, Iceland and Liechtenstein, says Norway’s Judin. He adds that the Norwegian authority doesn’t have any information about why the EU and EEA have been excluded from Bard’s launch.
By Wired, May 15, 2023
