European Union officials have issued Microsoft Corp. a legally binding request for information about two generative artificial intelligence features in Bing.

The European Commission, the EU’s executive branch, announced the move today. Microsoft has until May 27 to provide the requested information. Failing to do so could lead to potentially significant fines.

The request is part of the European Commission’s efforts to ensure tech giants comply with the bloc’s Digital Services Act, or DSA. The legislation requires large online platform operators such as Microsoft to monitor for several types of risks in their services and mitigate them. That includes risks associated with generative AI features.

In March, the EU asked Microsoft, Google LLC, Meta Platforms Inc. and several other tech firms to provide information about their respective generative AI features. The goal of the requests was to help officials determine whether the companies are complying with the DSA’s provisions about mitigating risks to users. The second request for information that the EU sent to Microsoft today was issued because the tech giant failed to share certain “internal documents and data” in its initial response.

The EU can impose fines equal to up to 1% of companies’ annual worldwide revenue if they submit incorrect, incomplete or misleading records in response to a request for information. Should Microsoft not disclose the data that officials asked for by the March 27 deadline, it could potentially face additional penalties. Failing to provide requested information can lead to fines equal to up to 1% of a company’s annual worldwide revenue as well as smaller periodical penalties.

The new request for information focuses on two generative AI features that Microsoft provides as part of Bing. The first is the search giant’s built-in Copilot chatbot, which is powered by OpenAI’s GPT-4 large language model. The other feature the EU is scrutinizing, Image Creator by Designer, allows users to generate images with natural language prompts.

The EU believes the two features may have breached the DSA “regarding specific risks stemming from Bing’s generative AI features.” The risks that officials have identified include the features’ potential to disseminate deepfakes and perform “automated manipulation of services that can mislead voters.” The EU is also concerned about hallucinations, AI responses that contain false or misleading information. 

Microsoft told TechCrunch in a statement that “we have been fully cooperating with the European Commission as part of the voluntary request for information and remain committed to responding to their questions and sharing more about our approach to digital safety and compliance with the DSA.”

Image: Microsoft