All of today’s generative AI is a beta – not just iOS 18

Mark Gurman’s latest newsletter suggests that Apple may choose to label the AI features in iOS 18 as a beta or preview, similar to the way the company labelled Siri when it launched.

Gurman’s piece has a lot of impressive details about Apple’s plans for iOS 18, from customizable icon colours to AI-generated personal emoji. But it also contains a criticism I feel is unwarranted …

The narrative that Apple is behind in AI

Gurman says:

But even now, there are signs that the company’s AI initiative is a work in progress. Apple is considering marketing the capabilities as a preview (at least in developer beta versions before a formal launch in September), indicating that the technology isn’t yet fully baked.

That would be an inauspicious move, especially when you consider Apple’s history. Siri launched as a beta test in 2011, and — given its struggles keeping up with the competition — arguably still feels like one.

This narrative – that Apple has been caught out, and is being left behind in AI – predates the explosion of generative AI into the mainstream. People have been saying for years that Siri is dumber than competing intelligent assistants like Alexa, and I’ve previously pointed out the “yes, but” rider to that claim.

Today it’s generative AI. Tech companies like OpenAI, Microsoft, and Google are all actively offering generative AI products and services; Apple hasn’t yet. Ergo, says much of the media, the iPhone maker has been caught napping.

This latest report plays into that idea. That even when Apple finally does move into generative AI, it will be with an early, tentative set of features it will label as a preview or beta.

I do think Gurman is very likely right that Apple is planning to use a beta label – but not to criticize the company for doing so. Because it seems to be there’s a simple truth too often ignored …

All generative AI is a beta

The world seems so caught up in the excitement of what generative AI can do that we seem to be willing to overlook its often egregious flaws.

I made this point when ChatGPT was new and shiny, pointing to examples of everything from demanding to be paid for translation work to calling an engineer and physicist a psychopath and a monster.

Some of those more dramatic issues may have since been fixed, but AI hallucinations – where a system confidently gives completely wrong or utterly bonkers answers – remain an everyday experience.

Just in the last few days the dietary advice from AI tools has included adding glue to pizza, eating rocks, and making spaghetti with gasoline.

Apple may just be more honest about it

So yes, I do believe that Apple is considering labelling the AI features in iOS 18 as a beta – because they are. Just like every other generative AI product of service.

Perhaps the final word on this should go to ChatGPT:

Outputs can sometimes be factually incorrect, nonsensical, or inconsistent […] AI models can inherit biases present in their training data, leading to biased or unfair outputs […]

In summary, generative AI can be reliable within certain parameters and use cases, but it is important to be aware of its limitations and actively work on improving and validating its outputs.

Sounds like a pretty good definition of a beta to me.

Photo by Solen Feyissa on Unsplash

FTC: We use income earning auto affiliate links. More.