Can Apple Intelligence break the pattern of half-baked AI?


Apple on Monday outlined how its new artificial intelligence features will let you ask a revamped Siri to play the latest episode of your favorite podcast, rewrite a hard-to-read message using the Mail app on your Mac, and even let AI generate customized images for group chats.

In theory, it all sounds useful and fun, but two realities make it hard to trust any of the AI ​​features Apple executives described in pre-recorded videos, especially since they weren’t demonstrated live and weren’t available for many journalists to try out.

First, other companies like Google, Amazon, and Microsoft boast similarly useful AI features that don’t quite work as well as hoped, and second, Apple has a shaky track record in AI, having shoved the dumb Siri around for over a decade.

If 2023 was the year that AI mania soared, 2024 should be the year to “prove it.”

Too many companies have promised magical, useful AI, but in reality, much of it hastily built and barely functional, not very helpful, insecure, or error-prone — like Google’s new AI-powered search that suggests you eat glue — are wasting time, energy, and trust in new forms of AI.

Apple’s new AI-powered features could be great. Most people won’t be able to take advantage of the Apple AI features until after September, and they may need to buy a new iPhone to use them.

What Apple’s AI demo showed (and didn’t show)

Apple outlined two layers of AI that will be built into the new iPhones, Macs and iPads: its own AI and OpenAI’s ChatGPT AI, which will take over for tasks that Siri and Apple’s apps can’t handle.

For instance, Apple indicated that if you need AI help to write a customized bedtime story for your child, your phone may ask for your permission to share a prompt written ChatGPT.

But I didn’t see why it wouldn’t be easier to just use the ChatGPT app or website to generate bedtime stories (either Apple’s own AI isn’t up to the task yet, or Apple didn’t want to risk its AI ruining children’s stories).

Get caught up in

Summarised stories to keep you up to date

Apple also lowered the bar for itself by not promising an all-around AI assistant that can take the steps out of planning a vacation or ordering groceries.

No company’s AI can do that, not even the all-encompassing AI assistants from Microsoft, Google, or OpenAI (which can’t even peer into all the places where relevant information lives, like your Gmail messages, Excel spreadsheets, and Expedia travel invoices).

Instead, Apple said its AI will make it smarter and easier to do the things people already do in its apps, and in future in other companies’ apps.

Some of the features Apple showed off, like AI-assisted writing and AI to edit objects from photos, are now standard features from Google, Microsoft and Samsung.

Even with the low bar, it’s not clear that Apple, the company behind Siri, the clunky iTunes that even its top executives derided, failed self-driving car software, and an autocorrect feature that thinks you’re typing “duck,” is capable of building a robust, flawless AI.

A pattern of over-promising and under-delivering

Many AIs promise magical things, but in reality they are deeply flawed or primarily useful for only a few tasks.

We don’t blame people who don’t use new forms of AI on a regular basis; most people do. A recent survey by the Reuters Institute for the Study of Journalism found that only about 7% of Americans use ChatGPT daily.

OpenAI introduced a more conversational version of its chatbot in May, but it didn’t have most of the features the company described. Google scaled back both its AI image generator and its AI-powered search results this year after high-profile missteps. Meta AI is a bit of a goof. Amazon unveiled an error-prone, AI-upgraded version of Alexa last year, but it’s yet to be released. (Amazon founder Jeff Bezos owns The Washington Post.)

These missteps aren’t Apple’s fault, and the company is typically cautious about releasing products that aren’t finished.

But Apple’s AI stellar performance comes at a time when the company and other companies are under immense pressure to demonstrate their AI expertise to investors, employees and business partners. It’s hard to know whether AI is truly built for users or primarily for corporate profit.



Source link

Related Posts

Next Post

Follow Us

Recommended

Instagram

    Please install/update and activate JNews Instagram plugin.

Highlights

Trending