Apple Explores Integration of Third-Party AI Models to Enhance Siri

The company has talked with both of the AI firms and asked them to train versions of their models that it could test on its cloud infrastructure, Bloomberg reported Monday (June 30), citing unnamed sources.

Apple is in the early stages of considering this move, hasn’t made a final decision, and is still actively developing in-house models for use with Siri, according to the report.

Apple did not immediately reply to PYMNTS’ request for comment.

The company currently powers most of its AI features with its own models and has been planning to use that technology for a new version of Siri that would be released in 2026, according to the report.

While Apple allows OpenAI’s ChatGPT to answer some web-based search queries in the voice assistant, Siri itself is powered by Apple, per the report.

Adopting third-party AI models could allow Apple to offer Siri features that would be competitive with the AI assistants available on Android smartphones, the report said.

It was reported June 12 that Apple aimed to bring an AI-powered upgrade of Siri to market in spring 2026, after facing delays and failing to meet its original goal of fall 2024.

The company originally announced the features it planned to add to Siri in June 2024. The capabilities were marketed last year and were set to be launched in fall 2024.

Its targeted release date was later shifted to spring 2025 and then, in March, postponed to sometime in the coming year.

Over that time, technical challenges led to Siri being rebuilt entirely and management responsibilities being shifted around within Apple.

It was reported June 8 that Apple had run into bugs during the process of trying to develop its own large language models over the machine learning technology that powers Siri.

On June 6, it was reported that the company’s stock price had seen its worst performance since at least 2010 as Apple faced a slew of problems, including its challenges with AI.