Microsoft Faces EU Fines for Incomplete AI Risk Data Submission
The European Union has issued a warning to Microsoft, indicating that the company could face a fine of up to 1% of its global annual turnover under the Digital Services Act (DSA). This comes after Microsoft failed to respond adequately to a request for information concerning the systemic risks posed by its generative AI tools. The EU’s request, made in March, sought detailed information about potential risks, particularly those related to Bing’s AI features such as “Copilot in Bing” and “Image Creator by Designer.”
The European Commission has emphasized the importance of this information, especially in light of upcoming elections and concerns about AI’s impact on civic discourse and electoral processes. The Commission has given Microsoft until May 27 to provide the required data or risk enforcement actions. Fines under the DSA for providing incorrect, incomplete, or misleading information can reach up to 1% of global annual revenue, which could amount to several billion euros for Microsoft, given its reported revenue of €195 billion for the fiscal year ending June 30, 2023.
The DSA imposes additional obligations on very large online search engines like Bing to mitigate systemic risks, including disinformation. Generative AI technologies are particularly scrutinized due to their potential to produce misleading information and deepfakes. The EU’s focus on these issues is heightened by the upcoming European Parliament elections, making the timely and accurate provision of information by Microsoft crucial.
Microsoft has stated its commitment to creating safe online experiences and cooperating with regulators. The company claims to have been fully cooperating with the European Commission and remains dedicated to addressing the Commission’s concerns. Microsoft also highlighted its efforts to prepare its tools for the 2024 elections and to safeguard voters, candidates, campaigns, and election authorities from AI-driven risks.
Source: EU warns Microsoft it could be fined billions over missing GenAI risk info | TechCrunch