Unlinked

My night in LinkedIn Jail, and how weird algorithms determine our fate.

Photo by engin akyurt on Unsplash

I recently had a scary encounter with the LinkedIn algorithm which one astute reader referred to as “Deeply Wrong writes itself”. The experience made me contemplate the role that LinkedIn plays in my life and the lives of millions, and how we are all at the mercy of the mathematical background rules that surround us.

Those of you who know me outside of SubStack are likely aware that I’ve been quite a prolific poster on LinkedIn for the past couple of years. Even though I’ve been on the platform for twenty years(!) at this point, it wasn’t until the winter of 2022 that I started posting semi-regularly because that’s when chatGPT 3.5 was released to the public. It was then that the topic of AI really heated up and I found myself deeply intrigued and concerned by the rapid hype machine that seemed to be overtaking any discussions grounded in reality.

Much like I do here with my two newsletters, Deeply Wrong and Humanware, on LinkedIn I share opinions about the darker side of technology as well as lessons learned from my career working as a software engineering leader. As a professional, I need to maintain my ability to earn a salary from my corporate experience, but on the other hand I’m someone who places a high value on my freedom to express what’s on my mind, challenge the status quo, push back on illegitimate authority, and ask what I consider to be important questions. So you’ll see me posting about the importance of being an empathetic leader one minute and then asking why we need artificial intelligence in our lives the next. It’s always been a risky juggling act, and I have definitely considered this dualism an act of self-sabotage at times. If I had a therapist I’m sure they’d agree.

I’m a lifelong advocate, user, and builder of technology, but choose to be very critical of arguably the most groundbreaking technology of my lifetime. As I’ve explained before, I consider myself a neo-Luddite, and the term Luddite should not be mistaken for one who is ignorant or afraid of technology but rather someone who approaches technology with the view that it is important to consider how technology is used and who wields it.

---

The crime

Cutting to the chase: on a Friday in October I did something as innocuous as re-sharing a post that someone in my network wrote. The post itself was a re-share of a research paper concerning the analysis of software development performed with the help of AI assistants which concluded that moderate use of AI resulted in a meager 4% productivity boost that was hard to measure or verify in any meaningful way. Basically it poo-pooed the idea that AI coding assistants can turn anyone into a hot-shot coding superstar overnight. I attached a few minimal lines of commentary that admittedly made the post a bit more spicy than the original:

If you're a (shitty) CTO who convinced your CEO it's time to reduce headcount on your tech teams because you believe, despite all practical evidence to the contrary, that AI copilots can take over most of the software development in your org, you need to read this.

Yeah, I was stirring the pot. Maybe I have some beef with leaders who never learned how to be human or value the humans working for them 🤔 One of the most dangerous and venomous consequences of AI snake oil is this notion that the technology is so good that it’s time to relegate human workers to the trashcan en masse. I can’t help but be emotionally fired up about this topic.

To my surprise, even though it was only a re-share, the post went viral. I’m sure I didn’t beat any records but this was huge for me: 100,000 impressions in 24 hours. The post had garnered 85 comments and been re-shared a further 60 times. It clearly hit a note with my network and extended audience. Even though my post was a re-share of a re-share, it seems to have captured even more attention than either of the preceding posts in the re-share chain. Another surprise was the more than 200 new followers and 59 new connection requests, some of which I approved the following day.

But on Sunday I woke up bleary-eyed, checked my email, and saw this subject line from an email apparently sent by LinkedIn:

“Jim, help others recognize you by adding a new profile photo”

I wasn’t sure what to make of this, since I already had a profile photo on LinkedIn that I liked and had used for about ten years. Reading on, it said:

Hi Jim,

We recently removed your profile photo because it does not appear to be a photo of you, which violates the LinkedIn Profile Photo Policy.

On LinkedIn, we require members to use their real identities to keep LinkedIn safe, trusted, and professional for all members. For this reason, we require member's profiles to show their real identity, including in their profile photo. Adding a profile photo with your actual likeness allows others to recognize you and helps create authentic and trusted connections on LinkedIn.

We recommend adding a new profile photo as soon as you can. When selecting a photo to upload, please remember that profile photos can get flagged for:

- Not being an image of you or an actual photograph
- Containing copyrighted material
- Containing content that's unauthorized for public distribution
- Being considered offensive

For more details about our policies, please visit Help Center

Thanks for using LinkedIn!
The LinkedIn Team

Well, that was an unusual email! And it made little sense. The profile photo I’d used for the past 10 years was one I was quite fond of. My wife took it when we were on a trip to my home country of England before the trials and tribulations of our second kid gave me grey hair. I looked like a real suave bastard. For once, I had a really good haircut that I liked, and I was genuinely happy as I posed for the photo. My wife had used her good camera, and it looked very professional. The location was inside a gorgeous oak-walled pub on a beautiful spring day in the Wiltshire countryside. I was drinking a deliciously crisp pint of Guinness (which was strategically cropped out of the photo before I uploaded it to LinkedIn.)

Befuddled by the email, I wanted to figure out what I was supposed to do to remedy whatever problem had occurred with my profile. The email was sent from LinkedIn’s no-reply address, so I couldn’t just respond to it like a normal person. It seems like they leave you to somehow work it out for yourself, so I clicked on the ‘Help Center’ link to find out more.

---

---

The punishment

Instead of ending up on the help center, the link routed me to the LinkedIn Login screen. Weird. It had been a long time since the last time I logged in. My credentials had been stored on this phone for 2 years. Accordingly, I entered my email and password, hit submit, and waited for the help center to load. Instead, I was presented with an even more shocking message, which is when I realized I’d been piledrived into social media oblivion, I was locked out:

Your account has been restricted. Please verify your identity. You may then appeal this decision.

Wow. What? So first they told me my profile photo was removed and that I should upload a new one, then without giving me any time to respond or fix whatever had gone wrong my account was now restricted. Hells bells. What the heck was going on?

So it was that I panicked and gave up all pretense of privacy on the internet and obediently went through the invasive identity verification process via Persona, a 3rd-party service that LinkedIn partners with. Using the mobile app, I was directed to look at the camera and maneuver my head so that it appeared inside a head-shaped frame drawn over my reflection in the camera. Then I had to turn my face side-to-side so that it could capture both profile views. I wondered if I would have to strip and scan my whole body next, or at least let a laser scan my eyeball. Next, I had to take a photo of the front and back of my driver's license, which I dutifully did without even thinking about it because of how nervous I was becoming about the prospect of losing my LinkedIn account. For some reason, it rejected my license, for reasons that it didn’t explain, so as a last resort I opted to scan my passport, which luckily seemed to do the trick. They thanked me for complying and asked me to wait up to 72 hours for “notice of your appeal.”

All this felt super uncomfortable. I still had no idea why my account had been restricted or what it had to do with the profile photo they’d initially removed. And all this despite paying $239 a year for LinkedIn Premium. At this point, I had a moment to reflect on all the ID verification steps I’d just gone through and it occurred to me that I’d been so quick to assume all this was legit that there was a possibility I’d just been totally scammed into giving up all my personally identifiable information to a total criminal!

Even after calming down and doing some after-the-fact due-diligence research into the Persona verification service I wondered, what if Persona itself is ever breached? How easy would it be for someone to clone my entire identity using whatever is stored on those servers from my official documents? How much trust can we place in any online entity no matter how much security and compliance measures they purport to have taken? Isn’t it only a matter of time before any such company is exploited? Why does a company like LinkedIn need to take such extreme measures to validate a user account. Can’t they see that I’ve been a user for 20 years and for the past 6 years I’ve been logged in on the same mobile device and my desktop has been in the same location?

But really what choice did I have? I considered that, for many reasons, I needed to be able to access LinkedIn. It’s critical to my livelihood. So I needed to do whatever was necessary to be allowed back in, even if it didn’t make a whole lot of sense. This is where I started to think about what LinkedIn means to me, to us, and how much of a monopoly it has over our careers.

---

Dependency

A lot has been said about how LinkedIn can be a cringe-fest or a haven for inappropriate virtue-signalling. There is even an entire subreddit devoted to how unprofessional and, at times, bizarre the so-called professional networking platform can be. Nonetheless, it is hard to argue with the fact that, for better or worse, LinkedIn has become a critical and perhaps unavoidable aspect of our professional lives. As one of the afore-linked Youtubers above attested: "Everybody has to be there but nobody wants to be there."

But social media is (to some extent) what you make it. Personally, my experience on LinkedIn has been quite enriching, at least in the past couple of years. For a long time I only used it as a repository for a resume that I barely needed to keep up-to-date since I had long tenures at just a few companies and rarely found myself between jobs. But after Facebook went south by aiding and abetting the proliferation of election distortion propaganda, and after being tired of fighting neo-Nazis on Twitter and then being permanently banned for saying something untoward about the proximity of Rand Paul’s face to his neighbor’s fist, I found that LinkedIn was a much more mature, civil and interesting place to hang out. Since chatGPT first entered the scene, I’ve cultivated a network of curious and like-minded PHD researchers, professors, data scientists, sociologists, futurists, psychologists, education experts, game designers, software engineers, technical leaders, writers, musicians, and artists who share an interest, if not disdain for the problematic aspects of AI and algorithmic bias, and I very much enjoy learning from and sharing with this niche community. I’ve crossed paths with my fair share of ‘LinkedIn Lunatics’, and probably a few people who think that sums me up too, but there’s always the block and mute buttons.

At the time of writing in 2024, LinkedIn is one of the longest-running social media companies and now boasts over 1 billion members across 200 countries [Source]. Most importantly, studies suggest the majority of employers use LinkedIn to search for job candidates, making it a crucial tool for anyone who wants to find gainful employment. It’s not just a social media platform: it’s the de-facto place to store your resume and establish your professional brand.

Consider being unemployed and looking for a new job without a LinkedIn profile. I can provide no evidence to prove that not having a LinkedIn profile will harm your ability to find new employment, but common sense suggests that your job search will be a lot harder if you don’t exist on the most commonly searched network used by hiring managers and recruiters the world over. For this reason, LinkedIn has been targeted with a class action lawsuit accusing it of monopolizing the concept of a professional network. If you lose your LinkedIn account or delete your profile, you’re essentially invisible to future employers.

This is why I felt an enormous sense of fear and anxiety when I realized that, for reasons that may never even be explained to me, I could be on the verge of permanently losing my LinkedIn account and all of the connections, followers, and content that I’d put work into over the past few years. Loss of status, loss of social connections, loss of community, loss of personal brand, a destabilizing effect on my professional identity: a social media site that is also the de facto bridge to the working world has an outsized impact on our sense of personal validity and belonging. It wasn’t always this way, and I strongly suspect we are just buying into a false illusion at the end of the day, but we have become addicted to the dopamine distribution algorithm, we have allowed sites like LinkedIn to dictate the terms of our social engagement, and in the case of LinkedIn it does have a material monopoly as the professional online network, so it’s incredibly hard to imagine a successful career and work life without it.

Even though I’m currently on a career break and not looking for a new employer, like many LinkedIn users I rely on my professional network for other things like building my personal brand, building up a network of potential leads, clients, or customers and as a subscription funnel to my SubStacks. Whatever I hope to achieve as a solopreneur in the future or if I’m ever in the market for another corporate job, I’m going to need LinkedIn. There’s just no foreseeable way to avoid this dependency.

I’m certainly not the only one who has experienced being put in LinkedIn jail, or “online gulag” for no apparent reason. A quick Google search of "LinkedIn account restricted" will return a legion of stories from people talking about their experience or offering various services or advice to help you get out of said jail. This is a very common occurrence.

---

Unconscious arbiters

LinkedIn is a private company and, like other private social media companies, is well within its rights to suspend or remove user accounts whenever it feels like a user has gone against the stated usage policies as laid out in the terms and conditions. Unfortunately, there is no law that says the company has to tell you why you are being reprimanded. Which is quite frightening in a kind of Kafka-esque way. Just like the character Joseph K in The Trial, the LinkedIn moderators — at the first level, just a soulless AI algorithm — can accuse you of some wrongdoing based on evidence that might never be presented to you, and then sentence you accordingly, with only a fleeting, nebulous chance of appeal. This is the psychological weaponry of a totalitarian entity. The AI has no conscience or feelings of remorse and is never wrong in the sense that a machine can never be right or wrong. If you’re lucky, a human moderator will eventually check your case and may allow you to appeal the decision, but since content moderation is terrible by design, I suspect that in many cases the AI remains the only judge, jury, and executioner and users are just deleted without trial. And there is nothing any of us can do about it.

As someone who has to exist in the corporate sphere to earn a living and provide for my family, the idea of being excluded from essential networking and job-seeking tools is somewhat terrifying, and like poor Joseph K, being monitored and judged by unknown, unseeable algorithms provokes feelings of extreme helplessness.

---

LinkedIn is just one example of an algorithm that holds sway over our lives. Weapons of Math Destruction by Cathy O’Neil, The Chaos Machine by Max Fisher and Unmasking AI by Dr Joy Buolamwini all detail the surprising ways in which algorithms dictate aspects of our lives, including what mortgage rate we receive when buying a home, the type of healthcare we receive when we are sick, how we are treated by law enforcement and the judicial system — even down to the length of sentencing we might receive — and how we are targeted by facial recognition systems and advertisers. Algorithms are everywhere and we can no longer avoid them or circumvent the influence they have on our lives. Not just small or temporary effects: algorithms contributed to the massive 2008 financial crisis which impoverished millions of Americans for years and caused a cascade of social and political issues we are still reeling from today.

A more subtle way that algorithms affect us on social media platforms is shadowbanning: where the things we post can be hidden from view without our knowledge because we have supposedly infringed upon some rule according to an over-simplifying machine that does not understand nuances such as a poster who speaks out against terrorism is not the same as a poster inciting terrorism, or how LGBTQ counterspeech is not the same as the hate speech that necessitated the response. I’ve seen so much of this since the attack on Israel in 2023, with some LinkedIn and X.com users seemingly targeted for suspension or removal for speaking out even though they called for peace, not violence. In some ways, the algorithms threaten us because of their intelligence and advanced capabilities, in other ways they threaten us because they are simply too dumb and do not understand the world we inhabit or the common sense rules and sensibilities that explain our behavior.

It’s disconcerting, to say the least. We place way too much faith in these automated systems. There are many in today’s society who seem to be okay with this surveillance capitalism, but I am not one of them. The general public has unwittingly become the product rather than the consumer of products, and it does not bode well for the species.

Of course, moderation is critical and should always be a built-in feature of any social network, but it’s fair to say that these tools often fall short and cause more harm than good in certain scenarios, at least for civil-minded social media users who are not actively trying to cause trouble or hurt anyone.

In a peculiar twist, there is evidence that social media content recommendation algorithms might actually be biased in favor of more polemic and controversial content, even while moderation algorithms penalize random users who actually haven’t done anything wrong. And despite disturbing statistics like “over 90% of women users on LinkedIn have received unwanted sexual advances” becoming public there is scant evidence that things are getting any better and some women are now leaving the platform in disgust or out of fear of reprisals from disgruntled men.

Clearly, companies like LinkedIn at least seem to want to address these issues but I personally have not seen much evidence that their AI or human intervention has been able to target bullying and harassment with effective outcomes. In the past, I have reported LinkedIn users who have attacked me personally in public or in DM’s, some of whom even tagged my employer in comments meant to get me into trouble at work or cause me to be fired from my job. To my knowledge, none of these users were ever reprimanded or removed from the platform. But woe is he who uses the wrong profile photo?

---

Mysterious gods

I genuinely feared I would lose my LinkedIn account forever. For me, the nail in the coffin was a friend on SubStack who told me that when he tried to search for me on LinkedIn my name now redirected to a page not found message. My account wasn’t just suspended, it was totally gone!

But my story has a happy ending. As it turns out, I was able to count on someone in my LinkedIn network to help me get out of LinkedIn jail. That someone was an employee at LinkedIn who was able to get the ball rolling on a quick human check of my situation that resulted in me regaining access to my account just over 24 hours later.

When I regained access, my profile photo was still missing, so I uploaded a new one. Was it really my profile photo that had triggered this whole process? Does the algorithm somehow care that I was pretending to be 10 years younger than I appear today? Is it possible that it “saw” a more recent photo of me from a LinkedIn Live event where I had a medium-length beard and then presumed that the short-bearded person in my profile was not really me but some kind of fake imposter? Why would it care so much? I'm not using my LinkedIn profile to board a plane. What about all the users who use an AI-generated avatar instead of a real photo? How are they not breaking the rules of authenticity? Was my profile photo even the real problem, or was that just some odd manifestation of a moderating algorithm gone awry? Was my viral post with 100,000 impressions just a coincidence or was it the real trigger for all this mysterious action?

On the left: my profile photo for the past 10 years. On the right: older and more bearded.

Why wouldn’t anyone at LinkedIn just explain to me in clear terms what was happening with my account? Too many users to bother? Too big to fail? If I hadn’t been able to leverage help on the inside, would I still be waiting to hear from them? Would I ever receive a response from a human worker at LinkedIn, or was I destined to be another lost soul who simply gets a robotic response saying I had been permanently deleted from the platform for reasons that will never be explained? Did we really sign up to be treated this way?

The legal answer is yes: we agreed to the terms and conditions. But were we really prepared for this future where we are nothing more than account numbers on a spreadsheet and a trove of consumer data to be mined by countless 3rd parties or excluded, deleted, and ignored at the drop of a hat, based on decisions made by a probabilistic machine that will never feel sorry for any harm that it causes?

I’d like to finish on an optimistic note, but I couldn’t find anything to suggest things will get better any time soon. On the contrary, so much of the content posted to LinkedIn is apparently now AI-generated that it seems before long the platform will be mostly content and comments that are entirely non-human. Perhaps the moderating algorithms will find it easier to police synthetic content that has never been edited by human hands and we can all stop pretending we were ever encouraged to be our authentic selves?

Until then, feel free to look me up on LinkedIn and maybe we’ll be friends, unless the omniscient eye of the LinkedIn machine somehow gazes upon this post and decides to slap me with another suspension, in which case you won’t be able to find me and it’ll be like I was never there.