Twitter’s nudges to do the right thing

Twitter has been testing ‘prompts’ for assorted things recently.

In 2020, they were testing 2 kinds of prompts. The first was article prompts. If you attempt to retweet a tweet that has an article, Twitter would prompt you asking you if you may want to read the article before retweeting it.

How would Twitter know you have not read it? This works only inside Twitter – if you have read the article outside of Twitter (by discovering it or clicking it from elsewhere), Twitter obviously wouldn’t know.

These prompts are non-intrusive – they do pop-up before you want to perform an action that previously was available without a prompt, but you can ignore the prompt and continue to retweet the article without reading it.

So, this was a useful prompt. According to Twitter’s own data, people opened articles 40% more often after seeing the prompt, and people opening articles before retweeting increased by 33%. This is good progress to ensure that people know and realize what they are sharing.

Twitter also tested a reply prompt in 2020. This goes one step ahead of the article prompt that was not about your tweet’s content and was only about your retweet’s attachment.

In reply prompt, the pop-up prompts you to reconsider a possibly offensive reply — these could be insults, strong or hateful language. Obviously, this is English-centric at least during the testing – Twitter can read, identify and prompt such alerts for English language content only.

Here, Twitter actually reads your reply in real-time when you hit ‘post’. Instead of complying with your action, it interjects you with this prompt. This may seem a bit more intrusive than the article prompt considering your words seem to be getting verified, but in essence, this is not very different from Gmail’s attachment prompt that checks if the word attachment is in your email copy and if there is no attachment, you get a prompt.

Twitter shared results of this prompt too – if prompted, 34% of users revised their initial reply or decided to not send that reply. After being prompted once, people composed, on average, 11% fewer offensive replies in the future!


The most interesting aspect of these prompts is that Twitter is taking on the role of nudging us (the users) to do the right thing.

In an ideal world, we should be doing this on our own – question ourselves if we have read the article that we saw only through the headline before retweeting it. Or have an internal conversation to question ourselves if the way we have articulated a reply is appropriate enough in terms of language.

But probably because we do not behave the way we are supposed to behave online, it falls on platforms like Twitter to try and nudge us towards the ideal behavior.

So, imagine – all platforms (including Facebook, WhatsApp, Instagram, LinkedIn etc.) can potentially prompt us to be better. But they have been hoping that we will behave better on our own. We clearly haven’t been.

But these nudges could make the platform seem like digital nannies, and lead us to believe that they distrust us by default.

Could the platform use more nuanced data to selectively offer these nudges to repeat offenders, perhaps?


The larger point worth considering here is how we treat what we say online. If the same expression was meant as an exam paper, an article to be published in the media, a letter/email to someone… we may be giving it more attention. But when it comes to what we say online, on social media platforms, probably because it seems so impulsive and fleeting (and usually very short), we do not consider the import of its effect.

What we say online and how we say it is very, very important to how people perceive us. This ‘people’ includes total strangers who have no context of who we actually are in real life.

Despite this, most of us do not think twice naturally before saying something online. This is largely because we do not see the ‘people’ who may read it (while we are alone with our laptop/phone). And the added irony: we say a LOT more online than on any other format. Probably because of that effortlessness and frequency, we seem to take our online utterances for granted. So the format that could do the most damage, on the platforms that we are most often actively vocal, is the one place that we take our utterances for granted!

Comments

comments