Just leave - NFHN Reader

I have previously written about why I don’t think Twitter is for me, as a scientist. As Elon Musk and his financial partners have bought out Twitter and taken the company private, I feel it is necessary to re-evaluate the role of Twitter in the scientific domain.

Twitter was a valuable platform for scientific discourse in the past

The primary benefits that I found from Twitter boiled down to its network effect. Many academics used Twitter, so having a Twitter account put you in the same online social space. This made it easy to share articles and wisdom, to see opinions of experts in other fields, and to learn from other colleagues by their successes and failures. These interactions can be eye-opening and informative. But they are also a small subset of all the communication that is necessary for science to succeed.

Science is, in general, slow and methodical. It takes time to be certain of anything; everything else is speculation. To turn speculation into certainties, we scientists need to communicate with each other; methods, data, results, interpretation, comparison, and discussion are all vital parts of our jobs. Each part begs for its own unique forum to increase communication and understanding. And in science, we have lots of different forums! There are:

troubleshooting discussions,
brainstorming sessions,
supervisor meetings,
work-in-progress meetings,
department meetings,
journal clubs,
chat rooms,
local conferences,
international conferences,
classes,
workshops,
seminars,
journals,
and so many more!

We communicate in so many different ways, with different intents for each of them. There were some unique discussions on Twitter that were hard to find, or impossible, elsewhere. But for the most part Twitter was useful for finding out about these other forums that were more useful than Twitter itself.

This is why so many scientists flocked to Twitter. There was valuable information to be found and it was easy to find.

Twitter encourages engagement at the cost of intellectual communication

Over time, two main factors changed on Twitter. The first was that more scientists joined Twitter. Whether you want it to or not, the dynamic of a conversation will change when you have more people participating. A 1-on-1 discussion is very different from one person speaking to an audience. The focused dialogue of experts in a shared field will be different from experts speaking to the public.

More scientists being on Twitter spurred discussions that didn’t exist before. But because of the number of these conversations that were happening, it was more difficult to find the tailored ones from before. Outsiders would be able to weigh in on topics they didn’t understand. Insiders would get distracted trying to address misunderstandings. The original conversation was diluted. Confusion of technical concepts was almost guaranteed in 280 characters.

Brevity may be the soul of wit, but what if you need more than that?

The second change was how Twitter decided what content to show you. With so many people on the platform, it was easy to be overwhelmed. You needed to filter this information somehow, and Twitter would love nothing more than to do that for you.

The most obvious result of Twitter choosing what to show you is that you’d see more content that drives “engagement”. Views, click-throughs, replies, likes, and every other website metric are prized possessions for a social media company. But do you learn anything when you’re continually being shown “big, if true” stories? Do you ever find what you’re looking for if you’re directed to wander in circles?

Coherent thought became rarer, still, when the context from one tweet to the next would change so drastically. This effect was amplified when Twitter interjected with tweets and ads it deemed necessary for you to see. Care to respond to this opinion removed from context or nuance? Would you like to comment on the meme making its way through your field? How would you like to see these ads mixed in with the actual content you want to see? Wait, what were you trying to do before you opened this website? Never mind, it’s onto the next tweet.

Despite what Twitter wanted to show you, not every important thought is valid in every context. Not every question is worth answering. You don’t have to participate in every fight you’re invited to.

This user experience made it hard to get what you, as a scientist, needed from the platform. There was still valuable information that people put out, of course. But it became surrounded by so much irrelevance that it was now hard to find. You had to jump through the hoops Twitter put in your way to find that valuable information.

So what if you don’t want to jump through those hoops anymore? What can you do? Independent online communities using Mastodon and Discourse for their forums have seen tremendous growth over the last few months. You can certainly leave Twitter to join one of these scientific communities. But Twitter won’t make it easy for you to jump ship if you want to keep your connections.

Twitter makes it hard for you to leave

Thanks to privacy laws, Twitter is required to let you download your personal data archive. Once you download and unzip this file, you can open the Your archive.html file to view your data in a web browser. The archive includes an interactive view of your data which is stored in the data/ folder.

A common format used for data exchange on the internet is JSON. At a first glance, your data appears to be in JSON format. But if you look more closely, you’ll see that the files have the .js extension. These aren’t JSON data files. They’re JavaScript program files. You can tell they’re JavaScript because each array is an assignment to a variable. The structure of following.js is:

//                       👇 this is the variable assignment
window.YTD.following.part0 = [
  {
     "following": {
       ...
     }
  },
  ...
]

You can’t just use standard JSON processing tools to handle this data unless you manually change it or write JavaScript code. And Twitter collects a lot of data, so you’d have to change a lot of it. Sure, this format is nice for Twitter. But these download archives aren’t meant for Twitter - they’re supposed to be for you. Worse yet, this data isn’t even useful to you in the form provided.

Imagine that you have your friends’ and family’s contact information in your phone’s contact book. If you get a new phone and want to transfer those contact details over, you can usually export that data. But what if the exported data was just a big list of people’s phone numbers? What if there were no names, no birthdays, and no email addresses? Is this export function actually useful for you, the person using it?

How useful would you find your personal data archive when you look at your Twitter following information and find this:

{
  "following": {
    "accountId" : "1298765418904117248",
    "userLink" : "https://twitter.com/intent/user?user_id=1298765418904117248"
  }
}

That’s it. No names, no handles, no relationships. Only numbers.

How are you supposed to know who is who? Only Twitter has that information and there are two ways of getting it:

Write code to interact with Twitter’s API, or
Click on links for every single person and record the information yourself

Both of these options require cooperation from Twitter. But this is supposed to be your data. Why is it being withheld from you?

You’re stuck in a situation where you can:

Manually collect that data using tools that aren’t meant to make that easy,
Take the easy path and just stay where you are, or
Abandon all that valuable data completely.

Twitter is holding your data hostage. This jeopardizes whatever benefits to your career you feel that data and platform provide. And the price you have to pay for that data is to never leave.

But if there is a cost to leaving, what is the cost of staying?

If you’re a medical doctor and a patient comes asking for medical advice, do you say out loud the first thing that comes to mind? You’ve probably seen all kinds of cases that match this patient’s symptoms, from totally benign to extremely traumatic. But if you mentioned the best and worse case scenarios, you will create uncertainty and cause unnecessary stress in your patient. Things that are inconsequential to say out loud with colleagues may be misinterpreted by your patient and cause them to act irrationally. It would be irresponsible for you speak like this in front of them.

However, do you, as that same medical doctor, need to wonder about those possibilities and think about how to rule them out? Absolutely - that’s your job. But you do that in consultation with other experts where the negative impacts of your work won’t bring harm to your patient. You do that thinking in private.

Each forum for scientific communication, above, serves a different purpose. Primarily, each forum provides focus and context for communicating. But just as importantly, each forum limits the reach of things that are easily be misunderstood by the public.

How many papers show the over-representation of European data in genome databases can lead to a misunderstanding of genetics in individuals of non-European ancestry? How many times must geneticists debunk the myth that these differences make one group of people genetically “inferior” or “superior” to another? How loudly must we state that the differences we find in human genomes accounts for such a small percentage of what makes us all human?

Twitter, in its absence of context and uniformity of design, is not good at limiting this reach. Worse, these types of traps are exactly the things that drive “engagement”, amplifying how likely it is to reach the people it shouldn’t.

When the COVID-19 pandemic began, scientists opined to an audience of thousands, often without regard for the consequences of this behaviour. There were important questions raised, like “where did the virus originate?”, “how is SARS-CoV-2 able to so effectively infect humans?”, and “this virus is airborne, how does that change our defense tactics?”. Many had to rethink what they would say now that they were thrust into the spotlight during a crisis.

Others, however, leaned into the uncertainty and not only said stupid things, but doubled down in the presence of overwhelming evidence against them. Some grossly downplayed the real risk that COVID-19 presented. Many continue to do so while decrying every public health measure that was implemented along the way. These scorned charlatans become fuel for the fire of contempt that is ever-burning in ignorant hearts.

Twitter spreads misinformation like this because it is so engaging. To a non-scientist, it is easy to confuse “this is what a scientist is thinking” with “this is what the best scientific evidence says”. This drivel stands in stark contrast to the slow and methodical work that is necessarily hidden from view to avoid the traps present in any field. But that rigorous work cannot compete with noise like this in an environment that yearns for “engagement” above all else.

It’s a tale as old as time. The truth needs to be so many things - lies need only be believable.

Twitter is making scientific communication more harmful

Gatekeeping scientific discussions to prevent public harm, then, becomes a case of moderation. But who’s to say how moderation decisions should be made if you aren’t aware of the consequences of those decisions? Who should even be making that decision? One of the world’s wealthiest men? A private tech company? Those who have no knowledge of your field? People who do not belong to your community? People who openly attack the principles your community stands for?

What should you do when someone says something that can harm others? What if nobody knows who this person is? What if that person is a police officer? Or a doctor? Or a tenured university professor? Or your Chief Medical Officer of Health? Or the leader of your country?

When does ignorance and a refusal to learn become negligence?

Content moderation in any form is hard, but it is necessary in some form to create useful spaces that people want to inhabit. It is necessary to prevent direct harm via threats and abuse, or indirect harm through disinformation campaigns. Twitter, like academia itself, has never been perfect at this. But at least there was internal debate, panels, and entire teams to think through their actions and consequences. Now there is one man lying about a moderation council.

None of this will be any easier for the scientific community going forward. Twitter is now abandoning the enforcement of its COVID-19 misinformation policies and reinstating accounts for individuals who actively spout hateful and anti-scientific ideas. Some journalists are working with Twitter to present the filtering of hateful accounts as if that were a bad thing. Employees responsible for the safety of people on the platform are resigning in protest of Twitter’s new directions.

These decisions will make the problems we already face more difficult. And now that Twitter is a private company, its decision makers will likely become even less accountable than they were before. Aside from effective scientific communication for ourselves and the public, what ethical concerns does this environment create?

If you, as a scientist, found yourself participating in a forum that:

made it more difficult to do your work,
stoked bigotry and harassment of your colleagues, and
dangerously promoted falsehoods over genuine scientific inquiry

Would you stay? What would it take to convince you to leave?

Each academic field has its own muddied history that all of us have to contend with. Medicine struggles with the pain it inflicts and its treatment of the mentally ill. Chemistry and physics struggle with their weapons of war and environmental destruction. Biology struggles with its roles in justifying racism and genocide. When will scientists’ participation in a forum that threatens health and scientific ideals be added to the list?

Conclusions

Ethical considerations in science are always necessary, lest science be co-opted for unethical ends. And while we will continue to debate these historical issues, it is always worth reflecting on the ethical concerns of the here and now.

What do you do when the places you like stop being useful and start hurting people? What happens when platforms you use don’t just allow, but actively promote, hateful individuals that denigrate your work? What do you do when more money than sense turns a place that you cultivated for years into a husk of its former self?

What do you do when people you want nothing to do with show you the door?

You just leave.