Why Twitter Should Verify More Users: Towards a More Human Web.
And being a finalist at SF’s 2018 Reverse Hackathon
This past weekend, I participated in SF’s Reverse Hackathon run by HackMentalHealth. I had been very excited to attend as I had an idea how to humanize the web: verified users.
History of verified users
Back in the day, Myspace was a land full of pseudonyms. Facebook came along and said that people had to use their real names. Not only did they say that, they enforced it. To get a Facebook account in the early days, you needed to have an email address assigned by an Ivy League university. The universities already had a vetting system, and Facebook leveraged it. When you interacted with someone on Facebook in those days, you knew it was who they said they were, because, it wasn’t so easy to cheat the system.
The former CEO of MySpace said that this is why Facebook actually beat them out: “Facebook’s killer feature was that it replicated the real world by forcing people to use their real names, whereas MySpace users used pseudonymous handles,” says Jones.
As Facebook grew, it relaxed this policy so it could grow faster. Not everyone attended a university that could verify an identity and attach it to an email and these people wanted to join Facebook. Nowadays, one can sign up to Facebook with just an email. Facebook tries to limit the number of fake accounts, yet the platform still has a tremendous amount of phony users.
How are the other companies in this space dealing with anonymity?
Twitter has no real name policy. One can create a semi-permanent pseudonymous handle, and an even less permanent pseudonymous name on top of it. The profile was restricted to a small photo, 140 characters, and a few links.
Google tried to have a real name policy and got a lot of negative flack for it —e.g., Google’s Real Names Policy Is Evil.
Reddit has pseudonyms and no real name policy, and 4chan is pretty much the opposite of a real name policy, where people don’t even have persistent pseudonyms. Both of these communities suffer from a significant amount of abusive and inappropriate content.
Disinformation and impersonation
With the US election in 2016, many of us became more aware of the prevalence of fake accounts and bots on some of our favorite platforms and how they impacted our lives. Hoaxes spread through YouTube. Twitter bots bullied celebrities and impersonated everyday humans. Comments sections were overridden with polarizing content, that upon closer inspection, were algorithmically generated.
Facebook, a pioneer of the real name policy, admitted in their Q1 2018 report that in that quarter alone, they removed 583 million fake accounts and 837 million for spam. If they have the 2.2 billion users they say they have, then that would be 26.5% of all users being fake, 38% for spam, and a combined total of 64.5% for the two causes. Now, that may be an exaggeration, and yet, without clear data from Facebook, we’re unsure how many people on Facebook are who they say they are.
This problem becomes more severe on Twitter. With no real name policy, short profiles, and short interactions, Twitter is ripe for fake accounts and bots. It has become such a challenge that there’s even a Lifehacker article about how to find out if there is a Twitter bot impersonating you.
So, why does this matter? An impersonation account could share things that you never would — porn sites, gambling schemes, racist statements, and more. Imagine that a Twitter account stole your identity, taking your photo, name, and exact profile text, and shared inappropriate links with the public internet. Imagine what would happen if you’re applying for a job and instead of seeing your real profile, the recruiters see this one. And the worst part is that you may never know that others saw the fake profile.
Been there, fixed that
This problem has already been fixed and it was already fixed by Twitter. In 2009, Tony La Russa, the former coach of the St. Louis Cardinals, sued Twitter over the accounts that were impersonating him and defaming him. As a result, Twitter created the famous blue check mark — Verified Profiles. If you were deemed an impersonation risk by Twitter, typically a celebrity or other public figure, then you could get your profile verified by Twitter verifying your identity. Instead of relying on a university to do this as Facebook did, Twitter decided to do it themselves.
Verified profiles soon spread to other social platforms such as Facebook, Google+, Quora, Instagram, Pinterest, and many others. Most seem to have an option these days.
How to get verified? Be famous.
Well, on most platforms, the process is not clear on who gets verified and how the verification happens. Some platforms even claim that you need 10,000 or 100,000 followers, which, ironically, may encourage users to get fake accounts to follow them, just so they can be verified.
You don’t have to be a celebrity to be verified (or impersonated)
Since impersonation is no longer for just the very famous (and probably never was), some platforms are offering verified profiles to us common folk. The dating app Bumble, in an effort to squash impersonations and bullying, introduced verified profiles in 2016.
Here’s how the verification process will work. Bumble will randomly show you one of 100 different poses — like someone taking a selfie while winking at the camera or holding up two fingers, etc. You then take a picture of yourself mimicking this pose to the company, and they make sure it matches up with the person in your profile pictures.
These pictures will be reviewed by a team of real people working in-house, and even though it sounds like a big workload the startup says they have the capacity to verify people “within minutes”.
Simple and effective.
Furthermore, there is a new company, Hooyu, who, for $5, will verify someone’s profile for you. It seems to be a straightforward process, where the user will receive a link to submit a photo of a government issued ID, a selfie through the app, and links to social media accounts. Then the people at Hooyu will verify the identity of the person and send the result to the company who requested the verification.
What should Twitter do?
Expand its verified profile option to all users. It could use Hooyu, a similar technology to Bumble, or it could create one in-house. This will give increased confidence to its users that they are interacting with real humans and it will give the public confidence that the person with this name actually said it.
It will also allow them to do this without even having to address the issue of bots directly. Many platforms have been fighting to get rid of bots, and I think it’s more challenging to do it this way. Imagine, there are 400 accounts of me on Twitter. Which one is the real me? You have to get rid of 399 to find the real me. OR, you could verify the real me, and then the other 399 don’t even matter. Heck, they could grow to be 999 and it still wouldn’t matter because you would know who I am.
In 2015, it was estimated that 0.05% of Twitter was verified. That means 99.95% was unverified. What if Twitter could shift that percentage? 10% verified / 90% unverified? 50% verified / 50% unverified? 80% verified / 20% unverified? They could get rid of their bot problem without having to even admit that they have a bot problem.
Oh, and perhaps they could make money doing this. Maybe they charge a flat fee of $10 to get a verified profile. Maybe they charge a subscription fee to have a verified profile and the option to only interact with verified profiles — i.e., only verified profiles can retweet you or mention you.
Goodbye ad fraud?
And that doesn’t even mention the advertising side of things. Right now, studies estimate that upwards of 30% of online advertising is fraudulent. In other words, $30 million out of every $100 million that advertisers spend may go to bots and other fraudulent accounts. By verifying more profiles, Twitter could differentiate themselves: the ads they serve go to real, verified humans.
Hello, numbers we can believe
What if by doing this, Twitter again starts a proliferation of verification across social platforms? How would this change things? YouTube could show how many unique human users have watched a video. Apple could display how many unique human users have downloaded an app. Facebook could tell investors how many unique human users are on their platform. And we can feel confident that people are who they say they are.
Thank you everyone who worked/volunteered with HackMentalHealth to put on this event. It gave me a platform to further develop and share an idea that had been brewing for so long in my soul. Thank you.
And especially you, Stephen Cognetta. At one point in the hack, you asked me how I was feeling and I said that I was struggling — I wanted to work with the cool people on my team and yet I felt guilty for trying to force them to go with the idea in which I so firmly believed. You told me that it was OK for me to work by myself and my day turned around from there. I’m grateful that you have put so much energy into this and you were there for me when I needed help. Thank you.
Original slides below:
