Why AI’s fake friends will bring us together – but not how the inventors hope

A much-maligned smart pendant is just the latest in a series of launches that seemed designed to make people argue with each other, writes Andrew Griffin

(Moviestore/Shutterstock)

The saturated, Black Mirror-esque advert that announced the arrival of Friend this week was in many ways an argument against itself. It depicted a device that users hang around their necks, which is always listening and then sends messages to people on their phone, with the aim of becoming a real friend.

At the end, one of the Friend’s owners is talking to what appears to be a prospective partner, who remarks that the Friend “goes everywhere with you”. As the awkward chemistry between the two begins to fizz over, the Friend’s friend reaches out to touch it, and seems to think better of it, deciding instead to stay in the moment.

Some people have attempted to use this as a gotcha against Avi Schiffmann, the 21-year-old Harvard dropout who founded Friend. But he has suggested that’s exactly the point, to bring people together.

He has managed to shrug off all kinds of criticism. Of the $2.5m (£1.9m) investment, for instance, he spent $1.8m on the domain Friend.com, bringing himself plenty of free mockery alongside it. And he has been attacked by a rival, Nik Shevchenko, who made his own wearable AI product called Friend and released a rap video making that point.

Amid all this criticism, something is working. The video has now been viewed 23 million times on Mr Schiffmann’s tweet alone.

Whatever the world thinks of the Friend – and it is mostly, seemingly not good – it appears to be the latest in what might be the decade’s defining tech product: the innovation that the world loves to hate.

Those products might always have existed, and the Google Glass might be the precursor to today’s mocked tech products. But recent years have seen a flurry of development of them: NFTs, cryptocurrencies, smart goggles and glasses, and much more besides.

Those are not all, necessarily, products that don’t have a use or deserve all the ire that they are subjected to. But they are products whose most notable characteristic is that there is a vast swathe of the world who do not like them, and might even argue that their very existence is evil.

And yet they still come. Until the NFT bubble collapsed, celebrities were pushing them on TV and a community had sprung up around them that hosted real-life events. Occasionally, that hatred seemed to spur them on: taking the idea that people don’t know what they need to the next level, occasionally supporters of those controversial products have argued that it is precisely because people don’t like the innovations that it is important to support them.

At the same time, the rate of really useful inventions has slowed. Until the release of ChatGPT and other exciting generative AI, perhaps the last tech product that excited the world was social media – and TikTok, the youngest of those mainstream social networks, began seven years ago.

Avi Schiffmann’s ‘Friend’ device appears to be the latest tech product that the world loves to hate (Friend)

It is perhaps because AI has somewhat broken that relative quiet that so many people – both real innovators and opportunistic grifters – have flooded into the space. We have seen all kinds of attempts to turn the exciting technology into a saleable product, and for the moment most of them do not appear to be there yet.

The Friend is not the first attempt to fold existing artificial intelligence tools into a devoted hardware item. Before came the Rabbit R1 and the Humane Pin, both of which promoted themselves as an entirely new kind of hardware for the artificial intelligence age.

It’s not the first of them to run into trouble, either. Both Rabbit and Humane were initially greeted with excitement about the fact that they could really be the future, but once reviewers got their hands on them it was clear that future still wasn’t here yet. The idea was perfect for generating excitement, but the actual fact of it failed to deliver.

Friend may yet suffer the same fate. Though Mr Schiffmann has shown off his own Friend’s messages at length on his Twitter account, it will not be available to the world until the first quarter of 2025, and the recent press blitz was not accompanied by any hands-on time or meaningful evidence that it actually works or is being shipped at all.

Indeed, Mr Schiffmann has sometimes spoken about the device as if it is as much an art project as it is a tech product. It may be that it doesn’t actually have to be released to make its point – that we are lonely, that AI could be a useful way of resolving that loneliness, and also (probably unintentionally) that the world is very cynical about any attempts to actually do so.

There appears to have sprung up a whole area of technology that is in fact about entering into a conversation about itself, as much as it is about convincing people to actually pay for a product. That was surely the great genius of NFTs: it was rare that anyone bought them, but for a brief time they were nonetheless ubiquitous – much more a result of the fact that they seemed a lightning rod for argument and disagreement about the way the world is going than anything else.

In that way, the Friend might indeed have helped in its way to solve loneliness: it brought much of the world together this week to laugh at it.