Apple’s AI Strategy Shifts as Siri May Get Powered by Anthropic or OpenAI

Apple’s AI plans are taking a big turn, as reported by Yahoo Finance. The company is now considering using AI from Anthropic or OpenAI to power Siri, instead of relying fully on its technology. If approved, this would be a major change in Apple’s AI strategy, as the company has long focused on developing its own AI tools.

Apple has asked both companies to test versions of their models, Claude and ChatGPT, on Apple’s private cloud. The move shows Apple’s growing urgency to catch up in the fast-moving world of generative AI for its Apple devices.

Apple Explores Outside Help to Upgrade Siri

Apple is testing whether outside models can boost Siri’s performance. According to reports, it has asked Anthropic and OpenAI to create custom versions of their AI models to run on Apple’s secure cloud servers. These tests are meant to compare their results against Apple’s models, known as Apple Foundation Models.

If Apple goes ahead with the switch, Siri could be powered by either Claude or ChatGPT as soon as next year. This would mark the first time Apple lets a core part of Siri rely on third-party AI, rather than internal systems.

Apple already allows ChatGPT to answer certain web-based Siri queries. But the main assistant still runs Apple’s technology. A full shift would be a big step forward, and an admission that its internal models still lack the desired performance.

This move could bring Siri closer to competitors like Google’s Gemini-powered features on Android. Apple’s change in approach mirrors Samsung’s strategy of integrating third-party AI under its brand.

Internal Tensions and the Push for Better AI

Apple’s leadership has changed how it manages Siri and its AI efforts. Siri is now under the control of Mike Rockwell and Craig Federighi. They replaced John Giannandrea, Apple’s AI chief, after delays and struggles in Siri development.

Rockwell, who also led the Vision Pro project, ran tests comparing Apple’s own AI with third-party options. After several rounds, he and his team found Anthropic’s Claude to be the best fit for Siri’s needs. This led Apple to start talks with Anthropic about a long-term partnership.

But this shift has caused concern inside Apple. Some engineers on the Foundation Models team are being held responsible for Siri’s slow progress. There are fears that if Apple relies on outside AI for Siri, it may later do the same for other features.

Some top engineers are already considering leaving. Companies like Meta and OpenAI are offering multi-million-dollar salaries, while Apple is known to pay much less. One senior Apple AI researcher, Tom Gunter, recently left the company after eight years.

Apple Focuses on Privacy and Long-term Goals

Even if Apple uses Claude or ChatGPT, it wants the models to run on its cloud servers. These servers are powered by Apple’s custom chips and offer stronger control over user privacy.

Other features in Apple’s generative AI roadmap still rely on in-house models that run on users’ devices. These are used for tasks like summarizing emails or creating Genmojis. Apple plans to open up these tools to developers later this year.

So far, Apple has not allowed developers to access its cloud-based models. One reason is that the cloud system cannot support a wide range of third-party features.

In the long run, Apple still wants to build its top-level AI models. But for now, it is open to using third-party solutions to improve Siri quickly. The leadership team sees this as a temporary step to stay competitive while internal systems improve.