Google has finally released its AI chatbot Bard in the European Union, after previously delaying the launch due to privacy concerns. The company said that it has addressed the concerns of the Irish Data Protection Commission (DPC), which is the lead regulator for Google’s data privacy practices in the EU.
Bard can be considered Google’s swift response to Open AI’s groundbreaking chat service ChatGPT. Although both chatbots provide a similar service, Bard differs from ChatGPT by sourcing its answers from the internet and providing the user with regular search results, which some regard as more accurate and specific. ChatGPT on the other hand relies on the data it was trained on, which has a knowledge cutoff up to September 2021.
However, Google’s innovative new service has previously met resistance when it comes to launching its operations in EU countries. Deputy Commissioner Graham Doyle of the DPC mentioned the service did not launch in the EU as was expected last month, as Google failed to provide sufficient notice to the regulator. According to Doyle, the tech giant did not provide any detailed briefing or data protection impact assessment.
Google has now provided the DPC with the requested information, and the DPC has cleared Bard for release in the EU. However, the DPC has also warned that Google will need to continue to monitor how Bard is used and to ensure that it complies with the EU’s data protection laws.
Bard’s privacy measures refined
In a reaction to Politico, Doyle elaborated that “Google have made a number of changes in advance of [the] launch, in particular increasing transparency and changes to controls for users”, further adding that “We will be continuing our engagement with Google in relation to Bard post-launch and Google have agreed to carrying out a review and providing a report to the DPC after three months of Bard becoming operational in the EU“.
Google’s Senior Product Director, Jack Krawczyk, announced that Google has enhanced Bard with new features to prioritise „transparency,“ „control,“ and „choice“ for users. These additions are supposed to enable users to understand how their information is used, opt outs for personalisation and model training, and manage saved or deleted conversations. The user interface now integrates a Bard „privacy hub“ and offers contextual information. Furthermore, the chatbot has been expanded to be accessible in over 40 languages, including Arabic, Chinese, Hindi, German, and Spanish.
In an answer to reporters, Krawczyk mentioned that Google had „very productive conversations with privacy regulators here in Europe,“ not only with the Irish privacy regulators, but „many others“ as well. „We expect it to be an ongoing dialogue,“ he said.
EU member states tighten the reins of AI-service privacy regulation
Google Bard is now available across 230 countries and territories worldwide, yet the service has struggled profoundly to gain its current access to European markets. This has emphasised the pressures and demands AI-services face in joining the noticeably strict EU market.
Earlier this year, ChatGPT faced a temporary ban in Italy. The Italian Data Protection authority (Garante per la Protezione dei Dati Personali) temporarily banned the use of ChatGPT, following a data breach which triggered a thorough investigation by the Garante that revealed Open AI lacked compliance to transparency obligations, the principle of lawfulness of processing, and conditions applicable to child protection.
After discussions with the Garante, these compliance issues have been addressed by OpenAI, and ChatGPT has been fully functional again in Italy since April 28th, although wider concerns remain regarding its overall EU data protection compliance.
ChatGPT has also been under investigation in several other EU member states such as Spain and Germany. European data protection agencies currently are scrutinising the various privacy issues that generative AI tools raise under the umbrella of the European Data Protection Board (EDPB).
US Tech companies continue to face compliance obstacles in the EU
The release of Bard in the EU is a significant development, as it shows that Google is committed to bringing its AI products to the European market.
However, the privacy concerns that led to the delay of Bard’s launch also show that there are still challenges that need to be overcome before AI can be used safely and securely in the EU.
In June, the Irish DPC pushed Google to postpone the launch of Bard in the EU, which serves as a recent example of U.S. tech companies holding back on introducing products in the region.
Earlier in May, Open AI CEO Sam Altman expressed a reluctance of remaining in the European market, noting that Open AI products might no longer be available in EU countries in the near future. According to him, it seemed unworkable to comply with the plethora of rules and regulations that the European Commission has laid upon generative AI-software developers.
Altman has since come back on his previous statement, claiming he has no plans to cease EU operations by the end of May. The OpenAI CEO tweeted: “Very productive week of conversations in Europe about how to best regulate AI! We are excited to continue to operate here and of course have no plans to leave.“
It remains to be seen how the situation for AI services in the European market will develop, especially in the face of the expected impact of the upcoming AI Act.