Amazon Fresh, the till-free grocery shop that uses “just walk out” technology, is closing all 19 of its stores in London, just under five years after opening its first outlet. If you’re unfamiliar, the premise of the store is that shoppers identify themselves at the entrance, walk in, select the items that they want, and then a combination of AI, sensors and computer vision determine the items in their basket and process an automatic payment via a customer’s Amazon account when they walk out.

If that sounds weird and disorienting then I can assure you – having visited an outlet out of pure curiosity and having left distressed – it is. Among the reasons given for the venture’s failure, from location choices to struggling to differentiate itself in the market, one financial analyst has suggested that till-less technology “always felt a little awkward”. When I visited I wasn’t totally clear on how to get in or, frankly, how to get out. A sense of panic overwhelmed me as I wondered if the sensors would process me changing my mind about an item and putting it back on the shelf, or charge me for it. Would I be prosecuted if, say, a large box of cereal blocked the sight of a tin of sardines and thus escaped the sensors?

Of course every store has CCTV equipment, but the idea that sensors and cameras could be connected to my phone and track every item I touched felt like big tech overreach, surveillance on steroids. That you could just walk out of a shop without pressing pay seemed strangely incongruous with the direction of other grocery stores. Around two years ago the big Sainsbury’s down the road installed scan-receipt-to-exit barriers, a technology I had first seen in Paris, and which has been rolled out to many other big supermarkets. It is truly a nightmare. Not only does it feel like you’re going through an airport when you’re just picking up a meal deal, but the scanner is repeatedly faulty, often resulting in a pile-up of people trying to exit.

Then there is the failure of self-scan checkouts. These tills were meant to save time, but that possibility immediately collapses once there’s an “unidentified item in the bagging area” or the overwhelmed shop assistant has to approve someone’s age.

You might then think the idea of a till-free checkout would be a relief. But if anything, when you’re made to feel so distrusted and burdened by inconvenience it feels far more like a setup. No till? Surely someone is waiting on the other side ready to bundle me into a police van over an unscanned pot of pesto pasta.

Mostly though, the failure of Amazon Fresh reveals that we are simply not ready for technology like this. It is the kind of futuristic development that you might have imagined would totally change the face of high street shopping, but shoppers have roundly rejected it. Like our reluctance to take up self-driving cars, it’s about a lack of trust in being totally at the whim of technology. Some stores have been able to win over the public – the Japanese casual wear brand Uniqlo’s self-checkout technology is pretty frictionless and genuinely loved. But even then, as a frequent Uniqlo shopper, while the convenience is nice it makes me feel strangely isolated.

View image in fullscreen

We need, and maybe even like, other people. Whether it’s grocery or clothes shopping, having a little chat or a flirt with a store assistant makes the experience. Recently, after a frustrating and failed attempt to find a suit for a wedding, I soothed myself by spending far too much money on a lovely knitted jumper at Drake’s on Savile Row. The shop assistant told me I looked good in it and, seeing how flustered I was, offered me an espresso. For that alone I’ll be back to blow more of my money.

Of course I don’t expect that treatment on the high street or in a grocery store, but I do find myself missing the small comments of “I love these crisps, my favourite” at a supermarket till. And queueing, though I’ll rue saying this during the post-work rush, is not all bad. One of my favourite things to do in a supermarket queue is peer into other shoppers’ baskets to make a guess about what kind of evening they’re having or what kind of life they live. If you can simply walk out you might save some time, but you’ll learn less about the people around you, while a computer gets to know it all.

• Jason Okundaye is an assistant newsletter editor and writer at the Guardian. He edits The Long Wave newsletter and is the author of Revolutionary Acts: Love & Brotherhood in Black Gay Britain

Before we ever knew of the term “fake news” or understood the disruptive power of digital technologies in our elections, Ireland was a canary in the coal mine.

During the final TV debate before the 2011 presidential election, a tweet making false allegations was read out live on air, effectively derailing the campaign of Independent candidate Seán Gallagher. Fast forward to this year’s presidential election. Even before nominations closed earlier this week, unsubstantiated claims online were already shaping the conversation. Unlike 2011, this wasn’t a single post but a campaign, carried out across multiple platforms, piped into the feeds of hundreds of thousands of voters.

This time round, it is Fianna Fáil candidate Jim Gavin grappling with the darker side of social media. And as it stands, the Electoral Commission, responsible for safeguarding our elections, is unable to compel platforms to operate with greater accountability.

Two weeks ago, a string of highly defamatory claims about Gavin began circulating online. They offered no evidence, only allegations tailor-made for the social media age, where sensationalism thrives and attention is rewarded, not accuracy.

The source of these claims was Kieran Kelly, an ex-fisherman from Helvick, Co Waterford, who now splits his time between Dubai, the US and Indonesia, where he led an ocean cleanup project. Kelly, who has a history of sharing conspiracy theories and is a self-described “Trump loyalist”, has ties to the Irish Freedom Party, a small far-right party active in Ireland since 2018. He spoke at its ardfheis in September 2023, endorsed party leader Hermann Kelly before the June 2024 European elections and did an interview with Kelly online earlier this month.

For weeks, from abroad, Kelly has conducted a campaign against Gavin, trolling him and coupling derogatory allegations with claims the candidate can’t be “trusted”, clearly seeking to influence public opinion. He labelled Gavin a “globalist” – a pejorative term common in right-wing circles and popularised by far-right conspiracy theorist Alex Jones. It is used to vilify public figures accused of undermining national identity and sovereignty in favour of liberal values and corporate interests.

When asked by The Irish Times for evidence for his smears about Gavin, Kelly deflected and promised more “damaging reports” before polling day. “I’m only getting started,” he warned in one post.

Kelly published the allegations across X, TikTok and Meta’s Facebook and Instagram platforms. After the initial wave, he returned days later with a second volley of posts. His content received thousands of comments and nearly a million views.

Despite clear violations of these platforms’ policies on harassment, abuse and privacy, the posts remained live. No fact-check labels were added, no clarification offered.

Recommender systems directed users towards the content; others, unaware of the specifics, turned to platforms’ inbuilt generative AI tools to fill them in on the allegations. This kind of content is bolstered by financial rewards offered to users for creating popular, sensationalist content. Blue-tick accounts spreading these allegations on X stood to benefit monetarily from the virality of their posts.

Finally, just over a week later, after Gavin’s campaign wrote to the platforms and spoke to the media to draw attention to the problem, action was taken. First Meta removed the posts and then TikTok. But why did it take so long for these “malicious smears”, as Gavin characterised them, to be addressed?

[ Anatomy of a fake story: How anti-immigration candidates spread false information to boost their profileOpens in new window ]

Online platforms have already played host to other claims seeking to target the integrity of the election. Claims about candidates being “barred” or “banned” from running, or allegations the process was rigged, are growing.

After Maria Steen failed to secure enough nominations to make it on to the ballot, much of the online commentary claimed she “was blocked because she posed a threat” or because Ireland is “not a democratic country”. In fact, her prospective candidacy was governed by the same rules that have been in place since 1937.

[ Fierce blame game as Maria Steen’s backers point fingers in wake of failed presidential campaignOpens in new window ]

Many of those targeting the legitimacy of the nomination process are using rhetoric initially popularised by Conor McGregor, the former MMA fighter who made headlines throughout the summer over his wish to run for the Áras and govern in the style of a US president. McGregor repeatedly attacked the legitimacy of the nomination process and accused Tánaiste Simon Harris of “tyrannically blocking the will of the people of Ireland” for allegedly “obstructing” his candidacy. McGregor never even formally pursued a nomination but his “campaign” nonetheless left a mark. A recent report from the European Digital Media Observatory Ireland in Dublin City University found that 58 per cent of McGregor’s posts during his campaign – shared with his 10 million followers – featured false claims about the presidency, the Constitution or Irish history.

The spread of misleading information during elections is nothing new. What’s different now is the speed, scale and potential impact it can achieve. The warning signs from other democracies are clear. Elections in countries around the world have faced serious threats from co-ordinated and prolonged campaigns aimed at manipulating public opinion.

[ ‘Deep fake’ image of David McWilliams used to dupe social media users, says commentatorOpens in new window ]

Gavin was correct to describe this episode as a “failure of our digital culture” and an “appalling feature of social media”. This could be an opportunity to use new powers conferred on the Electoral Commission allowing it to compel social media companies to address disinformation during elections. But the section of the Electoral Reform Act concerning those powers, which was passed in 2022, was never enacted. Delays have arisen from legal disputes with the European Commission and a coalition of tech firms who argue that the proposed Irish law goes too far. Effective powers for the Electoral Commission may not eliminate disinformation, but conferring them could provide the commission with clearer authority and tools for intervention. Because right now an imbalance exists, allowing wild and unsubstantiated claims to spread without scrutiny. One thing is certain: Gavin won’t be the last electoral candidate targeted by a vicious online campaign.

Ciarán O’Connor is a researcher and journalist who focuses on extremism and technology