The European Union has issued a caution to Microsoft, indicating that the company could face a fine amounting to as much as 1% of its worldwide yearly revenue under the Digital Services Act (DSA). This comes after Microsoft did not comply with a compulsory request for information (RFI) concerning its generative AI tools.
In March, the EU had solicited details from Microsoft and other major tech firms regarding the potential systemic risks associated with generative AI tools. Recently, the Commission reported that Microsoft had not submitted some of the requested documents.
Microsoft has been given a deadline of May 27 to provide the necessary information or face possible enforcement actions. Under the DSA, fines can reach up to 6% of a company’s global annual turnover, but providing false, incomplete, or misleading information in response to an RFI can lead to a separate fine of 1%. For Microsoft, this could translate into a fine of several billion dollars, considering the company’s reported revenue of $211.92 billion for the fiscal year ending on June 30, 2023.
The Commission is responsible for monitoring systemic risk obligations for larger platforms under the DSA. The warning to Microsoft underscores a range of potent enforcement measures that could impose more severe consequences than any damage to the company’s reputation for not furnishing the requested data.
The Commission has specifically pointed out a lack of information regarding the risks associated with Bing’s generative AI features, particularly emphasizing the AI assistant “Copilot in Bing” and the image creation tool “Image Creator by Designer.”
The European Union has expressed significant worries regarding the potential impact of certain tools on public dialogue and the integrity of elections.
The Commission has set a deadline of May 27 for Microsoft to submit the required information. Failure to comply could result in a penalty equivalent to 1% of the company’s yearly revenue. Additionally, if Microsoft does not meet this obligation, the Commission may enforce “recurring fines” that could amount to as much as 5% of Microsoft’s average daily revenue or its global annual turnover.
In April 2023, Bing was categorized as a “very large online search engine” (VLOSE) under the Digital Services Act (DSA), which imposes additional responsibilities on it to counteract systemic issues such as the spread of false information.
The DSA mandates that major platforms take action against the spread of disinformation, a directive that directly affects generative AI technologies. Despite known issues, such as the propensity of large language models (LLMs) to generate and present false information as if it were true, leading tech companies continue to integrate GenAI into their core platforms.
Furthermore, AI-driven tools for creating images have been criticized for generating outputs that can be racially prejudiced or dangerous, including deceptive deepfakes. With the EU’s upcoming election, there is heightened focus in Brussels on the threat of AI-driven political misinformation.
The Commission’s request for information stems from concerns that Bing might have violated the DSA due to issues associated with generative AI, including phenomena known as ‘hallucinations,’ the widespread sharing of deepfakes, and the automated tampering with services that could deceive voters, as stated in a press release.
The Digital Services Act (DSA) mandates that specific services, Bing included, must conduct thorough risk assessments and implement appropriate risk mitigation strategies (as per Articles 34 and 35 of the DSA). The Commission has pinpointed generative AI as a potential risk in its guidelines, particularly concerning the integrity of the electoral process for the forthcoming European Parliament elections in June.
In response to inquiries, a Microsoft representative issued a statement affirming the company’s dedication to fostering secure online environments and cooperating with regulatory bodies on this crucial issue.
Microsoft has stated its full cooperation with the European Commission’s voluntary information request and pledges to continue addressing their inquiries and providing further insights into its commitment to digital safety and adherence to the DSA. The company emphasizes its proactive measures across its various online services to identify and reduce potential risks. This includes specific preparations for the 2024 elections to protect voters, candidates, campaigns, and electoral authorities. Moreover, Microsoft intends to maintain its collaborative efforts with fellow industry leaders under the Tech Accord to counteract the deceptive application of AI in the 2024 Elections.