Medium Has AI Bots, Too Many! I mean way too many!

Read carefully and be diligent.

Real and fake Medium members circle around us, wanting money or influence. In this story, I’ll narrate what happened, how to identify cloned authors, and what this community can do to resist and reduce the traps.

I’m skimming through the Medium morning recommendations in my e-mail, a few days back. This, breakfast, and Youtube summarizes is my usual morning. The recommendations were all unfamiliar authors, but two options grabbed my attention. Hesitant before clicking, I plunge into a 4 minute read, feeling skeptical and hoping not to land on click-bait.

As a skeptic, I loath short form content. Too often, too full of garbage or water. I’m overly biased against AI sludge. For work, I design systems that unintentionally create my own AI-gen sludge, so don’t get me wrong.

But I was curious and Medium recommendations are much better than a few years ago. Following this concise and pointed article by Tobi the Audio Writer, I realized that the first article of interest had boldly named the second author. I’m very confused. In the article linked at the end, he claims that a fake Paul Rose tried to scam him. Clues to validate this were missing, but he was not wrong.

Coincidence? After reading the second article (okay, skimming), checking the user profile, and reading through comments, it’s clear that this is real Paul. If I’m wrong, this is incredible AI, but no, seriously, it’s legit. I left a comment to link Tobi’s article and shine some attention to my curiosity of a fake account. A few days later, a new fake Paul showed up, not the original.

Let’s move to the (many) fake Pauls…

First, this is the real Paul. I’m neither associated to Tobi nor Paul, but I’m a little jealous looking at the stats. Nice! Both articles are great FYI. Tobi’s story is a fun rollercoaster thrill.

Is Tobi’s claim real? I lack evidence to confirm.

Days later…a suspicious notification. It’s not a joke, but I had fun with the graphic. Can you spot the tiny clue that this is not real Paul?

Medium can mostly solve this by requiring verification of authors and changing the font. Print fonts have existed since…let me think. The printing press in 1440? In the 1400s, “Turks conquer Constantinople, end of the Byzantine empire, beginning of the Ottoman empire”, said Google the AI scholar. I imagine the importance of detail in a printed wanted poster as “Dead or Alive, $1, TONl with an EL, not TONI with an EYE, for the crime of fake influence and intentional identity confusion”.

Let’s talk about motives and ways to spot fake authors…

I assumed that I would forget and stop caring what happens next. Real Paul has serious adoration. This fake Paui was identified using an old trick. Is it case closed after informing Medium staff? Nope.

— Suppose that Medium adds verification and each writer must pay $1 with credit to verify the real human. — Suppose Medium fixes the confusing font. — Suppose Medium requires original profile images. — Suppose Medium can detect attempts to evade detection of forbidden terms. — Suppose Medium adds detection of what I call spoofers, evil twins, puppets, phish-fleets, pixel-pirates, sham-sharks, hoax-hogs, faux-festers…

AI, automation, and hackers are all around!

What happens to compromised verified favored accounts?

In 2021, 8 billion passwords became broadly leaked. The theft happened from prior years, but became packaged and broadly known just after the pandemic. In 2024, fk’n ‘Ahh! I mean look at these search results. I was about to talk about the prior leak and found this. There are just under 8 billion people alive. Hackers, please take a rest. I mean get arrested.

Conclusion

AI-gen and bots are fun for kids and hackers alike, but I take a strong stance regarding ethical engineering. The bot was caught. Let’s take a moment to remember “Paul with an EYE” in a moment of …So what’s new? Anyways.

RIP Paui

I didn’t get around to quantifying the number of bots/hacks or show all findings, but I’m curious if anyone has metrics to share. I occasionally provide feedback to Medium staff through a form because it honestly could be much better. Leading AI product development, I push my team the same way. This got under my skin though and is beyond what I can change with a few words, so I decided to make the feedback public for broader awareness. I’m not trying to ruffle feathers and get clicks/views. The CTA is simple: be diligent and aware. Use multi-factor authentication. Use those randomized passwords that you can’t figure out a way to remember. My point is…

Medium Has AI Bots, Too Many! I mean way too many!

Related

— Medium.com: Real Paul’s article — Medium.com: Tobi’s article — Medium.com:About MFA — Medium.com:More about MFA — Medium.com:Randomized passwords — Google.com: Chrome Password Gen and Storage — Google.com: MFA

Another one?