Despite significant problems with how Siri is designed, according to the New York Times, Apple is testing generative AI ideas that may eventually find their way into the virtual assistant.
At Apple’s annual AI summit last month, staff members received a briefing on the company’s massive language model and other AI capabilities. According to reports, Apple employees, including those on the Siri team, have been trying language-generation ideas “every week” due to the popularity of chatbots like ChatGPT.
According to the paper, these next-generation AI technologies have shown how voice assistants like Siri, Alexa, and others have squandered their advantage in the AI race. Particularly Siri has encountered numerous obstacles that have prevented substantive advancements. John Burkey, a former Apple engineer who worked on Siri and was tasked with enhancing it in 2014, told the New York Times that the voice assistant is based on “clunky code that required weeks to update with basic functionality.”
The “cumbersome design” made it extremely challenging for engineers to add new functions. For instance, Siri’s database is “one giant snowball.” It has a substantial collection of phrases in around two dozen different languages. Burkey continued, “It goes in one giant pile” when someone wants to add a term to Siri’s database.
For example, updating the data set to include new terms necessitates recreating the entire Siri database, which may take up to six weeks. It can take up to a year to add more difficult features like additional search tools.
Burkey thinks that because of this, Siri never had a chance to develop into a “creative assistant” like ChatGPT. This week, OpenAI unveiled GPT-4, their next-generation AI engine, enabling ChatGPT to provide even more sophisticated responses.
This week, DigiTimes claimed that major corporations, including Apple, Meta, and Amazon, had re-evaluated how the technology is produced in response to the increased interest in generative AI spurred by ChatGPT. These businesses reportedly want to prevent Microsoft from maintaining its dominance in AI. Particularly Apple and Tesla are rethinking their approach to AI.