Related Post
A software developer who was dumped by his girlfriend by text message was handed a brutal summary of her message by Apple’s new AI assistant before he had even had a chance to read the message.
Nick Spreen, from New York, was notified of the end of his relationship by a brusque summary of the text thanks to the new Apple Intelligence feature rolling out to iPhones in the near future.
Instead of showing him the message in full, he was given the deadpan synopsis of: ‘No longer in a relationship; wants belongings from apartment.’
Mr Spreen, an experienced software engineer who runs his own tech firm, shared a screenshot of the ruthless abstract on X, formerly Twitter, where it racked up millions of views.
He added in a follow-up tweet that the summary was genuine and happened on his birthday, October 9 – and started a whole new conversation about just how helpful AI is in summarising long-winded messages.
Tech worker Nick Spreen had the news of his breakup delivered to him brutally via Apple’s new artificial intelligence summaries
He shared a screenshot of the brusque summary on X, formerly Twitter, which stated: ‘No longer in a relationship; wants belongings from the apartment’
He was using a test version of the Apple Intelligence features that are due to launch on newer iPhones, iPads and Macs later this month
Other social media users have poked fun at the Apple Intelligence feature for reducing warm family moments to cold, harsh summaries
‘For anyone who’s wondered what an Apple Intelligence summary of a breakup text looks like,’ he wrote next to the screenshot of the text summary.
‘Yes, this was real; yes, it happened yesterday; yes, it was my birthday.’
Asked if it was a fair summary of the texts, he replied to one user: ‘It is.’
Mr Spreen would later describe the AI-powered text message summaries as ‘dystopian’ in an interview with Ars Technica, in which he compared the Apple summaries to having bad news delivered by a personal assistant.
‘It was something along the lines of ‘I can’t believe you just did that, we’re done, I want my stuff’. We had an argument in a bar and I got up and left, then she sent the text,’ he told the tech site.
He added that he had shared the screenshot as a point on the detachment that comes with AI-generated summaries rather than his ex, who he said had been ‘outraged’ by the adverse publicity.
‘I do feel like it added a level of distance to it that wasn’t a bad thing. Maybe a bit like a personal assistant who stays professional and has your back even in the most awful situations, but yeah, more than anything it felt unreal and dystopian.’
In a subsequent email exchange with MailOnline, Mr Spreen revealed that he actually has a master’s degree in artificial intelligence – and adds that he doesn’t necessarily believe that what his phone did was ‘bad’ in the grander scheme of things.
‘I think people love to hate artificial intelligence because currently at this point in time there’s so many things that suck about it,’ he said.
‘I think sometimes the hate goes a bit too far though, and I think the fact that you can now choose to read a summary of your iPhone notifications, before tapping on the summary in order to see the same old notifications you got before, isn’t that bad of a thing.’
The tech worker was testing an early version of the Apple Intelligence software that will launch on Apple devices including iPhones, iPads and Macs on October 28.
As well as summarising long text messages, the software will be able to help people with their grammar and spelling, generate images and custom emoji, and enhance Siri to better understand spoken requests.
It is compatible with the newest range of iPhone 16s, as well as the iPhone 15 Pro and Pro Max – with older devices unable to meet the tough technical demands required to run artificial intelligence models directly on a phone.
The latest generations of iPhones will be able to run the Apple Intelligence software – which can summarise texts, help with writing and even create new emoji on the fly
Previous text summaries generated by Google’s Gemini AI programme have prompted derision – and it continues to make baffling comparisons (as above)
AI summaries could represent the next evolution of dumping by text – a relatively new phenomenon that has become normalised as text messages replace phone-calls as the go-to communication method. They are still seen by many as a no-no.
Pop star Britney Spears made headlines when she told ex-husband Kevin Federline she was divorcing him via an SMS. But some see it as kinder than the modern day phenomenon of ‘ghosting’ – where communication ends altogether without warning.
Relationship coach James Preece said the rise of summaries was ‘not going to be a good thing’ because of the delicate balance people have to strike when they write text messages to ensure the right tone comes across.
‘Normal texts themselves are OK in certain situations, but it’s very hard to convey someone’s emotions from a message,’ he told The Times.
‘So I encourage my clients, for example, to use emojis. When you’re condensing them down even shorter, it’s going to be really, really blunt and direct, and it’s going to be open for misinterpretation.’
Summarising reams of text has become a cornerstone of artificial intelligence products online and in tech products like phones since the explosion in AI software such as ChatGPT, Google Gemini and Microsoft CoPilot.
But entrusting AI software – which uses ‘large language models’ that digest billions of pages of text to predict how sentences could be stitched together – is already provoking conversations about whether it can be trusted.
Google was mocked earlier this year as it rolled out AI-powered summaries in search results that could not distinguish between trusted sources of information and jokes or deliberately misleading information online.
Among the worst automated search summaries were a suggestion that people making pizza add PVC glue to their cheese sauces to make them stickier and that people should aim to eat at least one rock a day.
In the US, where a more advanced version of the software has been rolled out, dumbfounded social media users post summaries in which the AI has attempted to compare incompatible concepts, or gotten them wrong.
One shared a screenshot of their attempt to compare an air fryer with the Ottoman Empire.
‘The Breville Smart Oven Air Fryer is a countertop oven that can air fry, toast, bake and more, while the Ottoman Empire was a vast territory that controlled Constantinople and was a center of culture, art and science,’ Google deadpanned.
Another shared the unexpected result of their attempts to distinguish a sauce from a dressing.
Google’s AI assumed they meant a wound dressing, as opposed to the type drizzled on salad, and wrote: ‘The main difference between a sauce and a dressing is their purpose: sauces add flavour and texture to dishes, while dressings are used to protect wounds.’
Viral tweeter Mr Spreen said that he believes most people now realise AI ‘hallucinates’ a lot – a term for when AI generates something false and presents it as fact – and ‘produces a lot of garbage’.
He added: ‘AI only regurgitates things that people have said on the Internet before. The issues only begin when big companies like Google put their brand and trust behind something that AI created and I think that’s a real problem.’
