.png)
February 17, 2025
The development and integration of generative AI into everyday applications presents challenges for the tech industry. This is currently evident in the efforts of Apple and Amazon to equip their voice assistants, Siri and Alexa, with advanced AI features. Reports indicate delays and difficulties, highlighting the complexity of the technology.
Apple plans to comprehensively modernize Siri and equip it with new AI capabilities. These include improved access to user data, more precise app control, and the ability to capture the context of screen content. Originally planned for April, the introduction of these features could be postponed to May or later due to software bugs and technical problems. Internal tests have shown that the new functions are not yet working reliably. A Siri fully based on a large language model, capable of conducting human-like conversations, is not expected until next year.
Amazon is also struggling with integrating AI into Alexa. The launch of a premium version of Alexa with customizable personalities, call reminder functions, and the ability to place orders or call ride-sharing services is delayed. Although a limited demo version was presented in 2023, public availability has been repeatedly postponed due to inaccurate responses in tests. Amazon is particularly concerned about a potential loss of customer trust should the new Alexa version not function reliably. The premium features are now expected to be available no earlier than the end of March, although some of these functions are already integrated into the free version of Alexa.
The delays at Apple and Amazon highlight a fundamental problem: AI systems based on probabilities are difficult to integrate into everyday applications. The prediction of words and sentences by these systems is inherently unpredictable and can lead to unexpected and erroneous results. Apple recently had to disable an AI-powered notification feature due to incorrect message summaries. Chatbots and AI research systems also continue to struggle with the unpredictability of probability-based word prediction.
Apple's integration of ChatGPT as a standard feature raises questions. Does the company have higher internal standards for its own voice assistant, or is it struggling to match ChatGPT's capabilities – despite ChatGPT's documented errors? ChatGPT now offers live video analysis in addition to conversational functions, while Google's Gemini offers similar features and serves as the default assistant on many Android devices. The competitive pressure in the field of AI-powered voice assistants is high, and companies are under pressure to offer innovative and reliable solutions.
The development of reliable and powerful AI voice assistants is a complex process. The current challenges at Apple and Amazon show that the technology is not yet fully mature. It remains to be seen when the new features of Siri and Alexa will finally be available and how they will be received by users. The reliability and accuracy of the AI systems will be crucial to the success of these technologies.
Sources: Bloomberg Washington Post The Decoder Threads Twitter TechCrunch Tech Guru Guy Ars Technica AI Invest ResearchGate Financial Times