Google and Apple Battle EU Over AI Assistant Rules

The European Commission is turning up the heat on Apple and Google, arguing that their dominance in the smartphone market gives them an unfair advantage in the AI assistant race. With roughly 5 billion active Android phones and iPhones worldwide, according to Omdia, these tech giants control the gateways to next-generation AI tools. Now, regulators want to break that hold.

Under the Digital Markets Act, a set of EU regulations that took effect in 2023, the Commission is demanding that Google grant broader access to Android for third-party AI agents by July 2027. The goal is to ensure that virtual assistants like Google’s Gemini and Apple’s Siri aren’t the only prominent options on millions of devices. About 427 million iPhones and Android phones fall under EU oversight, making this a significant battleground for the future of AI.

Apple and Google are pushing back hard, warning that opening their platforms could create serious privacy and security vulnerabilities. Apple announced in June it would not launch its new Siri AI assistant on iPhones and iPads in the EU, citing the DMA as the reason. Google’s Kent Walker, president of global affairs, stated that granting external apps “sensitive and powerful device permissions” would be a major security risk. He also argued that sharing search data would “weaken citizen privacy, risk business trade secrets and endanger national security.”

Some experts, however, are skeptical of those motives. Calli Schroeder, senior counsel at the Electronic Privacy and Information Center, told CNN that tech companies’ privacy concerns should be taken “with a grain of salt.” She questioned whether these arguments would carry the same weight if the regulations didn’t threaten their business models.

The Commission’s mandate goes beyond just access. It wants EU Android users to be able to use voice commands to operate third-party AI agents and have those agents control apps just as Gemini does. Google must also begin sharing search data with other search engines and AI chatbots starting in January. This could allow companies like OpenAI, whose ChatGPT is already on about 30% of EU smartphones, to gain further traction.

Runar Bjorhovde, an Omdia analyst, highlighted the potential disruption: “The question is what the implications for Google or the consumer might be if the user can utilize ChatGPT to book an Uber, completely bypassing the need to engage with Android.”

Apple echoed similar privacy fears, claiming the EU’s requirements would force it to “give any virtual assistant direct access to users’ private data.” Apple said it proposed alternative solutions that were rejected but will continue working with regulators. A European Commission spokesperson countered that gatekeepers under the DMA “can implement appropriate privacy, security and integrity safeguards.”

The deeper issue is the privacy tradeoff inherent in AI. Giving apps deeper access to operating systems could allow them to siphon personal information, as Schroeder warned. Michael Stokes, senior vice president at Veilant, noted that a malfunctioning AI agent with access to messages, location, and microphones could “exercise the user’s delegated authority to do other things that maybe the user might not have wanted.”

Google argues that phone makers like Samsung would be responsible for vetting and managing those permissions. But the broader challenge remains: users may not fully understand the risks. “We also want to make sure that users are aware of the risks they’re taking when they engage with AI systems,” Schroeder said. “And I think that awareness is where we really need to catch up a little bit more.”