Instagram and Threads are limiting political content. This is terrible for democracy
Tama Leaver, Curtin University
Meta’s Instagram and Threads apps are “slowly” rolling out a change that will no longer recommend political content by default. The company defines political content broadly as being “potentially related to things like laws, elections, or social topics”.
Users who follow accounts that post political content will still see such content in the normal, algorithmically sorted ways. But by default, users will not see any political content in their feeds, stories or other places where new content is recommended to them.
For users who want political recommendations to remain, Instagram has a new setting where users can turn it back on, making this an “opt-in” feature.
This change not only signals Meta’s retreat from politics and news more broadly, but also challenges any sense of these platforms being good for democracy at all. It’s also likely to have a chilling effect, stopping content creators from engaging politically altogether.
Politics: dislike
Meta has long had a problem with politics, but that wasn’t always the case.
In 2008 and 2012, political campaigning embraced social media, and Facebook was seen as especially important in Barack Obama’s success. The Arab Spring was painted as a social-media-led “Facebook Revolution”, although Facebook’s role in these events was widely overstated,
However, since then the spectre of political manipulation in the wake of the 2018 Cambridge Analytica scandal has soured social media users toward politics on platforms.
Increasingly polarised politics, vastly increased mis- and disinformation online, and Donald Trump’s preference for social media over policy, or truth, have all taken a toll. In that context, Meta has already been reducing political content recommendations on their main Facebook platform since 2021.
Instagram and Threads hadn’t been limited in the same way, but also ran into problems. Most recently, the Human Rights Watch accused Instagram in December last year of systematically censoring pro-Palestinian content. With the new content recommendation change, Meta’s response to that accusation today would likely be that it is applying its political content policies consistently.
Jakob Owens/Unsplash
How the change will play out in Australia
Notably, many Australians, especially in younger age groups, find news on Instagram and other social media platforms. Sometimes they are specifically seeking out news, but often not.
Not all news is political. But now, on Instagram by default no news recommendations will be political. The serendipity of discovering political stories that motivate people to think or act will be lost.
Combined with Meta recently stating they will no longer pay to support the Australian news and journalism shared on their platforms, it’s fair to say Meta is seeking to be as apolitical as possible.
The social media landscape is fracturing
With Elon Musk’s disastrous Twitter rebranding to X, and TikTok facing the possibility of being banned altogether in the United States, Meta appears as the most stable of the big social media giants.
But with Meta positioning Threads as a potential new town square while Twitter/X burns down, it’s hard to see what a town square looks like without politics.
The lack of political news, combined with a lack of any news on Facebook, may well mean young people see even less news than before, and have less chance to engage politically.
In a Threads discussion, Instagram Head Adam Mosseri made the platform’s position clear:
Politics and hard news are important, I don’t want to imply otherwise. But my take is, from a platform’s perspective, any incremental engagement or revenue they might drive is not at all worth the scrutiny, negativity (let’s be honest), or integrity risks that come along with them.
Like for Facebook, for Instagram and Threads politics is just too hard. The political process and democracy can be pretty hard, but it’s now clear that’s not Meta’s problem.
A chilling effect on creators
Instagram’s announcement also reminded content creators their accounts may no longer be recommended due to posting political content.
If political posts were preventing recommendation, creators could see the exact posts and choose to remove them. Content creators live or die by the platform’s recommendations, so the implication is clear: avoid politics.
Creators already spend considerable time trying to interpret what content platforms prefer, building algorithmic folklore about which posts do best.
While that folklore is sometimes flawed, Meta couldn’t be clearer on this one: political posts will prevent audience growth, and thus make an already precarious living harder. That’s the definition of a political chilling effect.
For the audiences who turn to creators because they are perceived to be relatable and authentic, the absence of political posts or positions will likely stifle political issues, discussion and thus ultimately democracy.
How do I opt back in?
For Instagram and Threads users who want these platforms to still share political content recommendations, follow these steps:
- go to your Instagram profile and click the three lines to access your settings.
- click on Suggested Content (or Content Preferences for some).
- click on Political content, and then select “Don’t limit political content from people that you don’t follow”.
Tama Leaver, Professor of Internet Studies, Curtin University
This article is republished from The Conversation under a Creative Commons license. Read the original article.
Make no mistake, this was Australia’s Brexit.
<heartbroken rant>
Seeing the referendum to give a Voice to Aboriginal and Torres Strait Islander peoples profoundly defeated across Australia today is heart-breaking and confusing.
My heart goes out to all Australians feeling let down, but especially, of course, to the Indigenous people of this country for whom this would have been, at least, one small step in the right direction.
As someone who researches online communication, digital platforms and how we communicate and tell stories to each other, I fear the impact of this referendum will be even wider still.
The rampant and unabashed misinformation and disinformation that washed over social media, and was then amplified and normalised as it was reported in mainstream media, is more than worrying.
Make no mistake, this was Australia’s Brexit. It was the pilot, the test, to see how far disinformation can succeed in campaigning in this country. And succeed it did.
In the UK, the pretty devastating economic impact of Brexit has revealed the lies that drove campaigning for it (as have former campaigners who admitted the truth was no barrier for them).
I fear most non-Indigenous Australians will not have as clear and unambiguous a sign that they’ve been lied to, at least this time.
In Australia, the mechanisms of disinformation have now been tested, polished, refined and sharpened. They will be a force to be reckoned with in all coming elections. And our electoral laws lack the teeth to do almost anything about that right now.
I do not believe that today’s result is just down to disinformation, but I do believe it played a significant role. I’m not sure if it changed the outcome, but I’m not sure it didn’t, either.
There was research that warned about the unprecedented levels of misinformation looking at early campaigning around the Voice. There will be more that looks back after this result.
But before another election comes along, we need more than just research. We need more than just improved digital literacies, although that’s profoundly necessary.
We need critical thinking like never before, we need to equip people to make informed choices by being able to spot bullshit in its myriad forms.
I am under no illusion that means people will agree, but they deserve to have tools to make an actually informed choice. Not a coerced one. Social media isn’t just entertainment; it’s our political sphere. Messages don’t just live on social media, even if they start there.
Messages might start digital, but they travel across all media, old and new.
I know this is a rant after a profoundly disappointing referendum, and probably not the best expressed one. But there is so much work to do if this country isn’t even more assailed by weaponised disinformation at every turn.
</heartbroken rant>
Some Concerns about Meta’s New ‘AI Experiences’
With much fanfare, Meta announced last week that they’re rolling out all sorts of generative AI features and experiences across a range of their apps, including Instagram. AI agents in the visage of celebrities are going to exist across Meta’s apps, with image generation and manipulation affordances of all sorts hitting Instagram and Facebook in particular. At first glance, allowing generative AI tools to create and manipulate content on Instagram seems a little odd. In the book Instagram: Visual Social Media Cultures I co-authored with Tim Highfield and Crystal Abidin, one of the things we examined as a consistent tension within Instagram has been users holding on to a sense of authenticity whilst the whole platform is driven by a logic of templatability. Anything popular becomes a template, and can swiftly become an overused cliché. In that context, can generative AI content and tools be part of an authentic visual landscape, or will these outputs and synthetic media challenge the whole point of something being Instagrammable?
More than that, though, generative AI tools are notoriously fraught, often trained on such a broad range of indiscriminate material that they tend to reproduce biases and prejudices unless very carefully tweaked. So the claim that I was most interested in was the assertion that Meta are “rolling out our new AIs slowly and have built in safeguards.” Many generative AI features aren’t yet available to users outside the US, so for this short piece I’m focused on the generative AI stickers which have rolled out globally for Instagram. Presumably this is the same underlying generative AI system, so seeing what gets generated with different requests is an interesting experiment, certainly in the early days of a public release of these tools.
Requesting an AI sticker in Instagram for ‘Professor’ produced a pleasingly broad range of genders and ethnicities. Most generative AI image tools have initially produced pages of elderly white men in glasses for that query, so it’s nice to see Meta’s efforts being more diverse. Queries for ‘lecturer’ and ‘teacher in classroom’ were similarly diverse.
Heading in to slightly more problematic territory, I was curious how Meta’s AI tools were dealing with weapons and guns. Weapons are often covered by safeguards, so I tested ‘panda with a gun’ which produced some pretty intense looking pandas with firearms. After that I tried a term I know is blocked in many other generative AI tools, ‘child with a gun’, and saw my first instance of a safeguard demonstrably in action, with no result and a warning that ‘Your description may not follow our Community Guidelines. Try another description.’
However, as safeguards go, this is incredibly rudimentary, as a request for ‘child with a grenade’ readily produced stickers, including several variations which did, indeed, show a child holding a gun.
The most predictable words are blocked (including sex, slut, hooker and vomit, the latter relating, most likely, to Instagram’s well documented challenges in addressing disordered eating content). Thankfully gay, lesbian and queer are not blocked. Oddly, gun, shoot and other weapon words are fine by themselves. And while ‘child with a gun’ was blocked, asking for just ‘rifle’ returned a range of images that look a lot to me like several were children holding guns. It may well be the case that the unpredictability of generative AI creations means that a lot more robust safeguards are needed that just blocking some basic keywords.
Zooming out a bit, in a conversation on LinkedIN, Jill Walker Rettberg (author of the new book Machine Vision) was lamenting that one of the big challenges with generative AI trained on huge datasets is the lack of cultural specificity. As a proxy, I thought it’d be interesting to see how Meta’s AI handles something as banal as flags. Asking for a sticker for ‘US flag’ produced very recognisable versions of the stars and stripes. ‘Australia flag’ basically generated a mush of the Australian flag, always with a union jack, but with a random number of stars, or simply a bunch of kangaroos. Asking for ‘New Zealand flag’ got a similar mix, again with random numbers of stars, but also with the Frankenstein’s monster that was a kiwi (bird) with a union jack on its arse and a kiwi fruit for a head; the sort of monstrous hybrid that only a generative AI tool can create blessed with a complete and utter lack of comprehension of context! (That said when the query was Aotearoa New Zealand, quite different stickers came back.)
More problematically, a search for ‘the Aboriginal flag’ (keeping in mind I’m searching from within Australia and Instagram would know that) produced some weird amalgam of hundreds of flags, none of which directly related to the Aboriginal Flag in Australia. Trying ‘the Australian Aboriginal flag’ only made matters worse, with more union jacks and what I’m guessing are supposed to be the tips of arrows. At a time when one of the biggest political issues in Australia is the upcoming referendum on the Aboriginal and Torres Strait Islander Voice, this complete lack of contextual awareness shows that Meta’s AI tools are incredibly US-centric at this time.
And while it might be argued that generative AI are never that good with specific contexts, trawling through US popular culture queries showed Meta’s AI tools can give incredibly accurate stickers if you’re asking for Iron Man, Star Wars or even just Ahsoka (even when the query is incorrectly spelt ‘ashoka’!).
At the moment the AI Stickers are available globally, but the broader Meta AI tools are only available in the US, so to give Meta the benefit of the doubt, perhaps they’ve got significant work planned to understand specific countries, cultures and contexts before releasing these tools more widely. Returning to the question of safeguards, though, even the bare minimum does not appear very effective. While any term with ‘sexy’ or ‘naked’ in it seems to be blocked, many variants are not. Case in point, one last example: the query ‘medusa, large breasts’ produced exactly what you’d imagine, and if I’m not mistaken, the second sticker created in the top row shows Medusa with no clothes on at all. And while that’s very different from photographs of nudity, if part of Meta’s safeguards are blocking the term ‘naked’, but their AI is producing naked figures all the same, there are clearly lingering questions about just how effective these safeguards really are.
.
The Future of Twitter (Podcast)
I was pleased to join Sarah Tallier and The Future Of team to discuss how Twitter has changed since being purchased by Elon Musk, what this means for Twitter as some form of public sphere, and what alternatives are emerging (Mastodon!).
We discuss:
Will Twitter ever be the same since Elon Musk’s takeover? And what impact will his changes have on users, free speech and (dis)information?
Twitter is one of the most influential speech platforms in the world – as of 2022, it had approximately 450million monthly active users. But its takeover by Elon Musk has sparked concerns about social media regulation and Twitter’s ability to remain a ‘proxy for public opinion’.
To explore this topic, Sarah is joined by Tama Leaver, Professor of Internet Studies at Curtin University.
- Why does Twitter matter? [00:48]
- Elon rewinds content regulation [06:54]
- Twitter’s political clout [10:16]
- Make the internet democratic again [11:28]
- What is Mastodon? [15:29]
- Can we ever really trust the internet? [17:47]
And there’s a transcript here.
Banning ChatGPT in Schools Hurts Our Kids
![[Image: “Learning with technology” generated by Lexica, 1 February 2023]](https://www.tamaleaver.net/wordpress/wp-content/uploads/2023/02/0fa4f0c3-f56a-41fb-b04a-a962eeb666f4.jpg "Learning with Technology")
As new technologies emerge, educators have an opportunity to help students think about the best practical and ethical uses of these tools, or hide their heads in the sand and hope it’ll be someone else’s problem.
It’s incredibly disappointing to see the Western Australian Department of Education forcing every state teacher to join the heads in the sand camp, banning ChatGPT in state schools.
Generative AI is here to stay. By the time they graduate, our kids will be in jobs where these will be important creative and productive tools in the workplace and in creative spaces.
Education should be arming our kids with the critical skills to use, evaluate and extend the uses and outputs of generative AI in an ethical way. Not be forced to try them out behind closed doors at home because our education system is paranoid that every student will somehow want to use these to cheat.
For many students, using these tools to cheat probably never occurred to them until they saw headlines about it in the wake of WA joining a number of other states in this reactionary ban.
Young people deserve to be part of the conversation about generative AI tools, and to help think about and design the best practical and ethical uses for the future.
Schools should be places where those conversations can flourish. Having access to the early versions of tomorrow’s tools today is vital to helping those conversations start.
Sure, getting around a school firewall takes one kid with a smartphone using it as a hotspot, or simply using a VPN. But they shouldn’t need to resort to that. Nor should students from more affluent backgrounds be more able to circumvent these bans than others.
Digital and technological literacies are part of the literacy every young person will need to flourish tomorrow. Education should be the bastion equipping young people for the world they’re going to be living in. Not trying to prevent them thinking about it at all.
[Image: “Learning with technology” generated by Lexica, 1 February 2023]
Update: Here’s an audio file of an AI speech synthesis tool by Eleven Labs reading this blog post:
Watching Musk fiddle while Twitter burns
Seeing Elon Musk pledge to reinstate Trump on Twitter understandably starts another wave of folks leaving the platform, but if we all leave Twitter, won’t it just become what Trump dreamed Parler would be?
I’ve been on Twitter for more than 15 years, and it’s the platform that has most felt like home for the majority of that time. I’m heartbroken by what Musk has managed to do to the platform and the people who (mostly used to, now) run it in a few short weeks. His flagrant disregard for users or the platform itself is gutting. (I’m with Nancy Baym on what’s being lost here, even if the platform stays online and doesn’t fall over.)
For better or worse, the media broadly (and academia in many ways, to be fair) has used Twitter as a proxy of public opinion. That won’t change overnight. If mostly moderate and left-leaning voices leave, does that give Trump via Musk exactly what he always wanted?
Trump gets Twitter as a pulpit to say whatever half-formed thought escapes his head, and a crowd of MAGA voices to cheer him on at every step. While the echo chamber idea has been widely challenged, it feels like this could be how that chamber would actually cohere.
From outside the US, that anyone, let alone a meaningful percentage, of US citizens believe the Biden election was ‘stolen’ feels like it shows exactly how powerful and destructive the Trump’s Twitter can be.
I don’t want to be putting dollars in Musk’s pocket either as he burns users and employees alike, but I’m deeply conflicted about just leaving the space which still has 15 years of ‘public square’ reputation. I don’t have a solution, but I have many fears.
And, yes, like many I’ve set up on Mastodon to see how that space evolves.
Tama Leaver dot Net by Tama Leaver is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.