Why Apple's Craig Federighi Says AI Must Serve Humans

At WWDC 2026, Craig Federighi didn't bury his point in slides. He fired it like a flare across an industry racing to show off capabilities instead of asking who benefits.

'AI is an incredibly powerful technology,' he said, 'but some are charging forward for the sake of progress itself, with little clear attention to the people these systems are supposed to help.' Short sentence. Sharp observation.

Federighi's remarks read like a critique aimed directly at companies that have flooded the market with fast-moving AI products over the past two years — names like OpenAI, Google and Meta came under the implied fire. His counterproposal? A measured, privacy-centered approach where intelligence learns from the user's device and context rather than broadcasting personal data into the cloud.

He framed Apple's path as conservative by design, not timid. The idea is simple: make models work with the data that already lives on your phone, tailoring results to you without exposing your life to others. It's a privacy argument. It's also a user-experience argument. When AI knows you — locally and respectfully — it can be both useful and trustworthy.

Then came Siri. Federighi was blunt: the new Siri is not a mere chatbot. He described it as a conversational, integrated assistant that helps in the moment, stepping into tasks with context and continuity rather than offering empty, flashy demos. Think less novelty; more utility. Think of an assistant that finishes the sentence you were about to type, that surfaces the right app at the right time because it understands what you actually need.

The contrast couldn't be clearer: rapid-release, generalized models versus curated, on-device intelligence tuned to real people. Which path wins trust? Which wins market share? The tension is the story now. Developers and consumers will decide whether they prefer arms-length spectacle or quiet, practical intelligence that respects their data.

Whatever side of the debate you fall on, Federighi's message is a reminder: building clever systems is only half the job. The other half is ensuring those systems serve humans, not the other way around — and that question will shape the next chapter of AI.