Why I gave up trying to delete myself from the internet

An enjoyable trip down memory lane soon became a boring full-time job

This article was originally published on 8 November, 2024 in 1843 magazine.

 

Facebook remembers things about me that I wiped from my own memory a long time ago. Since I joined in 2006 I have been invited to 1,920 events, including the fifth birthday party of Movida, a long-defunct London nightclub; wake-like drinks to mark the start of the Brexit process; and a vegan conference (I have never been vegan). It also records my membership of a bewildering array of groups, including one for followers of the Rashtriya Swayamsevak Sangh (RSS), a Hindu-nationalist paramilitary organisation, and another for people who grew up in Huntington, West Virginia, once “America’s fattest city” (I am neither a Hindu-nationalist nor an obese American).

 

I wasn’t trawling through social media to reminisce about teenage nights out, or my more recent adventures as a foreign correspondent. Instead, I was on a mission to delete myself from the internet. My reasons were practical rather than ideological. I was about to host a podcast about India’s prime minister, Narendra Modi. Any reporter who wades into the debate on Modi’s high-growth agenda or Islamophobic comments risks becoming a target for online abuse.

 

Especially if they’re a woman. The International Federation of Journalists, a union, estimates that almost two-thirds of female reporters have endured online abuse. In South Asia, that share is undoubtedly higher: many female colleagues have been targeted by trolls, who publish their home addresses and send them death threats. One Indian investigative journalist told me someone had superimposed her face onto a woman in a pornographic video before circulating it online. Taking down as much personal information as I could seemed prudent.

 

By deciphering a man’s transaction history, I found out whom he was living with, holidaying with, going out for meals with and even buying drugs from

 

As a reporter I know how easy it is to find stuff out about people online. The other day I was researching a piece about the newly appointed chief executive of a Nasdaq-listed company. I messaged the people who had written endorsements on his LinkedIn page to ask what sort of manager he was. Public profiles on Venmo, a peer-to-peer payments app, are a goldmine. When they’re transferring money to a friend, people often add a message or a string of emojis. By deciphering a man’s transaction history, I found out whom he was living with, holidaying with, going out for meals with and even buying drugs from (it was either that or an eclectic order of medication, sweets and diamonds).

 

These experiences should have made me more careful about what I divulged online. They didn’t. More than once I’ve handed over my name and email address to log into “AddisFreeWifi123” on a layover in Ethiopia, called my mother to let her know I arrived safely, then logged into a banking app to transfer foreign currency to my overseas account. Until a couple of months ago I had never checked my privacy settings on social media.

 

With the release of the podcast imminent, I began trying to minimise my digital footprint. Step one was to figure out what was out there. Thinking like a stalker, I put my browser in private mode and googled my name. The first few pages of results were filled with stories I had written over the years – most of them before I joined The Economist, which, of course, has no bylines.

Next I downloaded my personal data from Facebook, Twitter, Instagram and LinkedIn. It took days rather than minutes to come through, which I found reassuring, as if every inch of the sites had been scoured. Twitter sent me dozens of folders, one of which included every image I have ever shared via direct message. (These included several photographs of boarding passes, which brought home just how much time I waste picking fights with airlines over delayed flights.)

 

I was particularly worried about trolls finding photographs of me, so I put my name into Google Images. The search engine served up only images where my name was in the accompanying text, such as my photo in The Economist’s media directory. Fearing there could be others out there, I turned to PimEyes, a facial-recognition tool powered by artificial intelligence (AI). I paid £15 ($20) and uploaded a passport photograph. PimEyes searched the open web for images that matched (social media and video platforms aren’t currently searchable).

 

PimEyes spat out 26 photographs. There I was at my cousin’s wedding in 2014 – the photographers had posted the images on their website. And there I was in the background of a photograph of a well-known American senator, thrusting a voice recorder in her face as she walked the corridors of Capitol Hill (I spent the summer of 2017 reporting on politics in Washington, DC).

 

I enjoyed looking at photos from birthday parties, hen parties, weddings and baby showers. I marvelled at the number of men I know who have completed the Arch 2 Arc triathlon between London and Paris

 

One of the photos was of somebody else. It was from a Mexican news site, and showed a woman with dark skin and dark hair plotting her next move in a local chess tournament. It turns out that AI models, trained largely on images of white men, struggle to tell brown women apart.

 

Because I live in Britain, I have more control over my data than most. The General Data Protection Regulation, which was implemented by the European Union in 2018, is widely considered to be the toughest data-privacy law in the world. It dictates that people must be able to demand a copy of the information held on them, and ask for it to be corrected or deleted. Britain kept those rules in place after Brexit. A few other jurisdictions, including the state of California, have introduced similar laws in recent years.

 

In practice, getting information taken down is not easy. As an experiment, I chose a random webpage with my name on it – an article I wrote as a student for a newspaper in Florence, Italy – and filled in a form asking Google to remove it from searches. I acknowledged that the article wasn’t lewd or violent, but it included a few personal details about where I went to university and what I studied. Three days later my request was approved, but there was a catch. Google could delist the link only from searches in Britain and Europe. To get the article wiped completely I would need to contact the website itself.

 

It struck me that anyone could plug a photo of me into PimEyes. The site says its tool is only for personal use, but it is tough to police that sort of thing. To have my images removed from PimEyes searches I had to hand over a photograph, email address and a redacted copy of my passport (they reassured me they wouldn’t store any of this information).

 

I decided not to delete my social-media accounts, not least because they come in handy for reporting. Instead, I started deleting posts that revealed too much about me. At first, scrolling through my feeds was a fun, nostalgic process. I enjoyed looking at photos from birthday parties, hen parties, weddings and baby showers. I marvelled at the number of men I know who have completed the Arch 2 Arc triathlon between London and Paris. But by the time I got to 2020, I ran out of steam. It turns out deleting yourself is a full-time job.

 

This is when I started looking for professional advice. Rob Shavell, the co-founder of DeleteMe, a company that promises to do exactly what its name suggests, told me I had been worrying about the wrong things all along. “What you can see and delete is the tip of the iceberg,” he told me.

 

Shavell told me to stop and think before I allow an app to track my location (free apps should ring alarm bells – they have to make money somehow).

 

Shavell introduced me to the nefarious world of data brokers. These businesses buy data from apps and websites, organise it into databases, then sell it on. They scoop up data that people share without even realising it. Shavell told me to stop and think before I allow an app to track my location (free apps should ring alarm bells – they have to make money somehow).

 

It is possible to get data brokers to delete information. But I would have to contact them individually. By Shavell’s estimations, it would take roughly 100 hours to cover every major broker. And it wouldn’t be a one-off task. These firms refresh their databases every few months. Of course, DeleteMe could take care of it. But the company charges $129 per year, which is more than I could expect my editor to approve in expenses.

 

Having scaled down my ambition to remove every trace of me from the internet, I set aside an hour to take some basic precautions. On LinkedIn I opted out of personalised ads. On Facebook and Instagram I deleted the most cringeworthy posts. I wiped my search history on any platform that would let me. I set up two-factor authentication, so anyone trying to get into my accounts would require two separate credentials.

 

I had assumed most online trolls were unemployed men with little to do. Some, it turns out, are tech bros with well-paid jobs in Menlo Park, California

 

In the end, the backlash to the podcast was modest compared with what some other journalists have to put up with, but it was still depressing. Quite a few people sent me racist tweets and angry emails. What surprised me was the number of abusive messages I got on LinkedIn, an ostensibly professional network. I had assumed most online trolls were unemployed men with little to do. Some, it turns out, are tech bros with well-paid jobs in Menlo Park, California.

 

My advice to other people who are worried about their digital footprint? Unless you’re a spy, a crook or incredibly sensitive, please google yourself, check your privacy settings and delete social-media posts that make you shudder. And if you ever need to use the Wi-Fi at Addis Ababa Bole International, you can be jamesbond@gmail.com and log in just fine.