After a decade or more, the auto-remediation “failures” may be waning. Apple’s much-maligned spell-checking software is being upgraded with artificial intelligence. Using a sophisticated language model, the new autocorrect can not only match words against dictionaries, but also consider the context of words in sentences.In theory, it doesn’t suggest comfort when you mean integrationBecause it turns out that those words are not interchangeable..
Next-gen AutoCorrect is one of several small updates to the iPhone experience that Apple announced earlier this month. The photo app will now be able to distinguish between your dog and other dogs, automatically recognizing your puppy the same way it recognizes people who frequently appear in your photos. And over time, AirPods will get smarter about adjusting ambient noise based on what you listen to.
It may not be obvious from Apple’s description, but all of these features are powered by AI. At the conference announcing the update, AI, is now a buzzword for tech companies of all kinds. Instead, Apple used more technical terms such as: machine learning again transformer language model. Apple has been silent about the technology, and has been accused of being so silent that it is lagging behind. Sure, ChatGPT can make a half-hearted business proposal, but Siri can only set your morning alarm and not much else.but apple teeth It’s pushing AI in small ways, a gradualist approach, but it could still be the future for this technology.
Since ChatGPT debuted last fall, technology leaders have been less cautious about the potential of AI, for better or worse. OpenAI CEO Sam Altman said: tweeted Last month, he said AI was “the most amazing tool ever built.” Microsoft founder Bill Gates called AI “the most important technological advancement since the graphical user interface.” At a Google conference, Alphabet CEO Sundar Pichai spoke about “AI” 27 times in a 15-minute speech. (He also famously said that AI would be “deeper” than fire.)
Apple, on the other hand, isn’t even pretending to give the big picture when it comes to AI. John Gruber, a longtime Apple follower of his who runs the tech blog Daring Fireball, told me he doesn’t expect the machine learning features Apple announced this year to significantly change the experience of his iPhone users. rice field. They only nominally improve it. “We expect autocorrection to work,” he told me via email. “We notice when we don’t.”
The new AutoCorrect, available with iOS upgrades later this year, is like the less powerful ChatGPT in your pocket. Apple says the software will be better able to tweak it to the way we type and predict what words and phrases we’ll use next. When you ask ChatGPT, you access the same giant language model stored in the cloud that everyone else accesses. But much smaller, more personalized language models that power autocorrection will live on. on iPhone. Apple didn’t provide details on how the feature works, and it’s not clear the exact technical approach Apple is using here, said Tatsunori Hashimoto, a computer scientist at Stanford University. told me. Researchers, including Hashimoto, have worked hard to find ways to shrink large language models to fit his mobile devices.
AirPods, on the other hand, will now use “adaptive audio” to analyze the sounds around them and adjust accordingly. For example, your Airpods may automatically lower the volume of your music when you start talking to a barista at a coffee shop, and raise it when you stop talking. Apple says it uses machine learning to understand your overall volume settings to optimize your listening experience.
All of this is very Apple-ish, Gruber said, focusing on what the feature does rather than how it works. “The fact that you’re using AI behind the scenes is as important to users as, say, what programming language you used to create it,” Gruber said. It also highlights user privacy, which Apple has long prioritized (or at least claimed to prioritize). The company uses an “on-device” model, which may pose fewer privacy risks than giant cloud-based models like ChatGPT. “User data never leaves the phone, so it’s private in a way.” [and] That user-tailored model never leaves the user’s phone,” Mohsen Bayati, an AI expert at Stanford Business School, told me.
Some of the differences between Apple’s approach to AI and that of other technology companies can be explained by their respective business models. Not all tech giants make money the same way. Google and Meta control about half of the digital advertising market, and AI-powered chatbots could be another way to get us to buy things. Microsoft isn’t much involved in the advertising business, but it hopes adding chatbot capabilities to search will give Google a bit of a cut. Amazon’s huge cloud hosting business stands to benefit from the adoption of language models at scale (they have to exist somewhere!). Apple is a luxury brand and, above all, deeply committed to making using computers and mobile phones enjoyable. “It is therefore not surprising that Apple is approaching AI cautiously, with a product-oriented focus,” Gruber said.
Still, it’s still unclear how chatbots will appear in our daily lives, so the iPhone may be the first time many people encounter new advances in AI. ChatGPT was a hit, and within two months of its launch he had 100 million users, though it’s unclear how many of them still use it regularly. (An OpenAI spokesperson declined to give figures when asked about the current average number of monthly users.) Many companies are adding their own chatbot capabilities. The Instacart app uses AI to serve recipes, and Salesforce recently debuted something called “Einstein GPT.” —but chatbots still have practical limitations. They’re always making things up, they’re bigoted, they’re a copyright nightmare.
Small technical inconveniences can be frustrating, but there’s a reason why “No, I didn’t mean ‘ducking’, autocorrect” has become such a meme. AutoCorrect improvements apply to billions of mobile phones, tablets and computers. Currently, the majority of smartphones in the United States are iPhones, and the company counts he has over 2 billion active his devices worldwide. Other tech giants are also using his AI to upgrade existing products on a small scale. Google recently announced the ability to draft Gmail replies via chatbots. A scenario in which human-like chatbots take over is not the only way AI can change the world. Many small tweaks in front of you can turn out to be big. In a way, they already have. For years, machine learning has served us targeted ads, filtered our social media feeds, and helped determine search results.
Like the rest of Silicon Valley, Apple may soon make even bigger changes. Wedbush Securities tech analyst Daniel Ives (it’s Ive, not Ive) believes Apple’s new AI features are just “the appetizer before the main dish.” His team estimates the company has spent between $8 billion and $10 billion on AI over the past four to five years, the same amount Microsoft invested in OpenAI in January, and Apple’s AI is reportedly looking for talent.
Hey Siri, maybe your life is running out.
