A senior official said that the European Union sees a “lot of convergence” with Japan in thinking about generative artificial intelligence (AI). “Generative AI is where we see the most alignment,” European Commission Vice-President for Values and Transparency Vera Jourova told Reuters in an interview. Generative AI is the most advanced AI that can create content like text, code, images, or videos. These can be used for various purposes, from customer service chatbots to making nonconsensual deepfake porn.
The EU has been at the forefront of regulating emerging technology, with its hard-hitting AI Act laying down rules companies can be liable for. Other countries have offered mainly guidelines, but the EU’s law has more teeth than others because it allows for a court case to be filed against companies violating the regulations. “If it gets passed, it will set the standard for the rest of the world,” said Kris Shrishak, a technologist and senior fellow at the Irish Council for Civil Liberties.
A new EU-Japan report on security policy cooperation notes that both countries have a similar approach to the problem of trusting and regulating AI. However, the two have divergent levels of collaboration on specific dimensions of the issue. For example, high levels of coordination in threat perception correlate with an overlap between EU and Japanese domestic response. In contrast, medium levels of convergence indicate a lack of agreement on the seriousness of threats or differences in policy responses.
The report says that to address this discrepancy, the EU and Japan will work together to promote a framework for “trustworthy and inclusive AI” that will be applicable internationally. This will include a code of conduct, standards, and recommendations that governments, regulators, and industry could use to set principles for trustworthy AI. It will also seek to foster greater transparency in the use of AI, including by requiring companies to regularly report on their AI systems and submit them to third-party auditing.
In the short term, the EU and Japan will focus on building a joint action plan that will allow for greater collaboration on risk assessment, data protection, and public awareness of AI. It will also explore how to bring together their respective expert networks for joint initiatives such as training and research.
But it will take time for the EU and Japan to bridge a more compatible center of regulatory ambition, especially with the U.S. on a different timeline for developing its own rules on AI. It has already taken a while for the EU to develop its law on the groundwork laid down by the draft version of its AI Act, which is expected to be finalized by the end of this year and come into effect in 2024. This has put the EU further behind the U.S. and China in the global race to regulate AI. It may even need to catch up in the future, with other countries likely to adopt its stricter guidelines.
