Apple engineers allegedly testing AI-generated language features

AppleInsider may earn an affiliate commission on purchases made through links on our site.

With ChatGPT growing ever more popular, Apple has reportedly decided to step up its game and focus on natural-language search processing.

In February, Apple held its annual AI summit, an internal event that briefs employees on its Machine Learning and AI advancements.

The event was somewhat more subdued than expected and was most noteworthy for being the first in-person event that Apple had held in years.

However, the New York Times spoke to those familiar with the event, which seemingly offered more insight. According to the sources, engineers — including those said to be from Apple’s Siri team — are actively testing language-generating concepts.

Generative language would be a massive undertaking for Apple. Currently, Siri needs to be fed a prompt that already exists inside its database. So, if a user asks Siri a question that hasn’t already been integrated into its code, it responds that it cannot help the user.

John Burkey, a former Apple engineer, told the New York Times that upgrading Siri’s data set required engineers to rebuild the entire database — a task that could take up to six weeks. He believes that adding more complex features could take a year and doesn’t believe that Siri would ever become a creative assistant like ChatGPT.

More broadly, the New York Times discusses how tech companies are now expected to keep up with ChatGPT, with major players like Google, Amazon, and Microsoft rushing to integrate chatbot features into their products.

Microsoft has already begun testing its ChatGPT features, which have been integrated into its Bing & Edge products.

Google is bringing AI tools to users in Google Workspace and opening up its AI language model called PaLM for developers and businesses.

It’s unlikely that generative language is going away anytime soon. On Tuesday, OpenAI released GPT-4, an upgraded language model for ChatGPT that is more accurate than its predecessors.