3 minute read
CEO Apple Inc. (NASDAQ:AAPL) Tim Cookexpressed uncertainty about the company’s ability to completely prevent artificial intelligence (AI) hallucinations.
In an interview with The Washington Post published Tuesday, Cook acknowledged that Apple’s new Intelligence system could generate false or misleading information, The Verge reported.
Despite Apple’s efforts to ensure the quality of the technology, Cook acknowledged that the potential for errors still exists.
“I think we’ve done everything we can, including thinking very deeply about preparing the technology in the areas where we use it,” he said, adding: “So I’m confident it will be of very high quality. . But I would say, frankly, that it is less than 100%. I would never say it’s 100%.”
Apple Intelligence system presented at the exhibition Worldwide Developers Conference (WWDC) has introduced several artificial intelligence features for devices such as iPhone, iPad and Mac. These features will allow users to perform tasks such as generating email responses, creating custom emojis, and summarizing text.
However, as with other AI systems, there is a risk of generating false information, as shown in the AI model above. Twins from Google And ChatGPT from OpenAI.
At WWDC 2024, Apple also announced a partnership with OpenAI to integrate ChatGPT into Siri. Cook said Apple chose OpenAI because of its strict privacy measures and the quality of its model. He also hinted at possible future partnerships with other companies.
Apple Senior Vice President of Software Engineering Craig Federighialso hinted that the tech giant may use Google Gemini in the future.
You can also read: 5 stocks in focus on Wednesday: AVGO, ORCL, PLAY, MAMA, OXM
The problem of AI hallucinations is not new to the tech industry. Previously in the podcast episode Lex Friedman, Yann LeCunchief researcher at AI Meta Platforms Inc. (NASDAQ:META), explained the phenomenon of AI hallucinations.
“With autoregressive prediction, every time the AI creates a token or a word, there is a certain level of probability that that word will take you out of the set of reasonable responses,” LeCun explained, showing how chatbots can go astray in a conversation.
Last April, CEO Alphabet Inc. (NASDAQ:GOOG) (NASDAQ:GOOGL), Sundar Pichai, also admitted that artificial intelligence technology still struggles with “hallucination problems” without a clear explanation. “No one in this field has yet solved the problem of hallucinations. “This is a problem with all models,” he said at the time.
Image via Shutterstock
You can also read: Kramer: Eli Lilly will earn $1 billion from its new drug
For more information on this topic, activate Benzinga Spain alerts or follow us on our social media: X and Facebook.
Receive exclusive information on market movements 30 minutes before other traders
With a 14-day free trial of Benzinga Pro, available in English, you get access to exclusive information so you can get actionable trading signals before millions of other traders. Click here to start your free trial.
Is it true that electric cars pay less taxes? There are many tax incentives surrounding…
Driven by the idea that “time together is the best gift,” the brand is launching…
Another week in which we have a lot to think about in Spanish football.. A…
Adele watched the latest performance from her Weekends With Adele residency in Las Vegas and…
Mike Waltz speaks during a hearing at the Capitol in Washington on July 22, 2024…
between 60 and 80% of children hospitalized for influenza had no previous health problemsas warned…