Tim Cook, CEO of Apple, just admitted that he has made tremendous efforts to make sure that their new AI technology called Apple Intelligence is ready, yet there are no loopholes to show that it will be completely without mistakes or “hallucinations”. During an interview with The Washington Post, Cook said, “We have done all we can do, with much contemplation about the readiness of the technology in the area of its application; I think this will be very good stuff. But I must say frankly that is not all of it. There is no way I would say it’s 100% complete.”
The Apple Intelligence program was presented in the 2024 Worldwide Developers Conference also referred to as WWDC. It was intended for the purpose of incorporating cutting-edge AI features into iPhones and other Apple hardware. Apple wanted to give the point of emphasis on Personal Data Context and Privacy and so called it as Personal Intelligence while keeping itself away from AI or words related to AI. However, the specific Apple Intelligence is more framed under generative AI and may at some point cause hallucination.
These are clinically referred to as AI hallucinations where the AI system provides wrong or non-real content. These systems generate outputs conforming to the templates learned during training using the training data and in the process of doing so, can provide seemingly sensible but semantically wrong answers. For instance, Google’s Gemini AI has started uttering nonsense, including that cats have been to the moon, or that you should use glue to set cheese to pizza. Such concerns could relate to Apple’s AI and be problematic for some of the company’s businesses.
At WWDC 2024, the company also announced collaboration with OpenAI to incorporate ChatGPT into its devices, which could also enhance the practicality of Siri in handling of multifaceted tasks or providing searched information. Tim Cook highlighted the successful efforts of AI model by OpenAI for privacy, While he noted that his company may also adopt other AI models like Google Gemini in the future. This has however been put into confirmation by Craig Federighi an Apple Senior Vice President of Software Engineering during the post keynote.
Also read: Apple Ecosystem to Get Major Software Updates at WWDC 2024, No New Hardware Expected
Cook referred to Apple Intelligence as the new chapter in Apple innovation in his words made after the WWDC was through. The Apple deal with OpenAI seems like a trial in nature as Apple needs more time to tune its native models. It is probably one of the reasons why it is initiating it in beta as the intention is to use it and check the kind of response that it will receive.
Google’s CEO Sundar Pichai has defended AI hallucinations saying they are good for creative and art fields but practical uses have shown big problems. So, Google has reduced its AI overviews. Apple knows these problems so they are developing AI carefully. We’ll see how Apple Intelligence will do and how well it will balance personalization for each user and keep user’s data safe when used at scale.
For more such news, tech-news.in
