Content moderation is a hot topic in social media circles at present, as Elon Musk goes about reforming Twitter, while simultaneously publishing past moderation actions, as an illustration of how social media apps have gained too much power to control certain discussions.

But despite Musk highlighting perceived flaws in process, the question now is, how do you fix it? If content decisions can’t be trusted in the hands of, effectively, small teams of execs in charge of the platforms themselves, then what’s the alternative?

Meta’s experiment with a panel of external experts has, in general, been a success, but even then, its Oversight Board can’t adjudicate on every content decision, and Meta still comes under heavy criticism for perceived censorship and bias, despite this alternative means of appeal.

At some level, some element of decision-making will inevitably fall on platform management, unless another pathway can be conceived.

Could alternative feeds, based on personal preferences, be another way to address such?

Some platforms are looking into this. As reported by The Washington Post, TikTok’s currently exploring a concept that it’s calling ‘Content Levels’, in an effort to keep ‘mature’ content from appearing in younger viewers’ feeds.

TikTok has come under increasingly scrutiny on this front, particularly in regards to dangerous challenge trends, which have seen some youngsters killed as a result of participating in risky acts.

Elon Musk has also touted a similar content control approach as part of his broader vision for ‘Twitter 2.0’.

In Musk’s variation, users would self-classify their tweets as they upload them, with readers then also able to also apply their own maturity rating, of sorts, to help shift potentially harmful content into a separate category.

The end result in both cases would mean that users would then be able to select from different levels of experience in the app – from ‘safe’, which would filter out the more extreme comments and discussions, to ‘unfiltered’ (Musk would probably go with ‘hardcore’), which would give you the full experience.

Which sounds interesting, in theory - but in reality, would users actually self-classify their tweets, and would they get these ratings correct often enough to make it a viable option for this type of filtering?

Of course, the platform could implement punishments for not classifying, or failing to classify your tweets correctly. Maybe, for repeat offenders, all of their tweets get automatically filtered into the more extreme segmentation, while others can get maximum audience reach by having their content displayed in both, or all streams.

It would require more manual work for users, in selecting a classification within the composition process, but maybe that could alleviate some concerns?

But then again, this still wouldn’t stop social platforms from being used to amplify hate speech, and fuel dangerous movements.

In most cases where Twitter, or other social apps, have been moved to censor users, it’s been because of the threat of harm, not because people are necessarily offended by the comments made.

For example, when former President Donald Trump posted:

The concern wasn’t so much that people would be affronted by his ‘when the looting starts, the shooting starts’ comment, the concern was more that Trump’s supporters could take this as, essentially, a license to kill, with the President effectively endorsing the use of deadly force to deter looters.

Social platforms, logically, don’t want their tools to be used to spread potential harm in this way, and in this respect, self-censorship or selecting a maturity rating for your posts, won’t solve that key issue, it’ll just hide such comments from users who choose not to see it.

In other words, it’s more obfuscation than improved security - but many seem to believe that the core problem is not that people are saying, and want to say such things online, but that others are offended by such.

That’s not the issue, and while hiding potentially offensive material could have some value in reducing exposure, particularly, in the case of TikTok, for younger audiences, it’s still not going to stop people from using the massive reach potential of social apps to spread hate and dangerous calls to action, that can indeed lead to real-world harm.

In essence, it’s a piecemeal offering, a dilution of responsibility that will have some impact, in some cases, but won’t address the core responsibility for social platforms to ensure that the tools and systems that they’ve created are not used for dangerous purpose.

Because they are, and they will continue to be. Social platforms have been used to fuel civil unrest, political uprisings, riots, military coups and more.

Just this week, new legal action was launched against Meta for allowing ‘violent and hateful posts in Ethiopia to flourish on Facebook, inflaming the country’s bloody civil war’. The lawsuit is suing for $2 billion in damages for victims of the resulting violence.

It’s not just about political opinions that you disagree with, social media platforms can be used to fuel real, dangerous movements.

In such cases, no amount of self-certification is likely to help – there will always be some onus on the platforms to set the rules, in order to ensure that these types of worst-case scenarios are being addressed.

That, or the rules need to be set at a higher level, by governments and agencies designed to measure the impact of such, and act accordingly.

But in the end, the core issue here is not about social platforms allowing people to say what they want, and share what they like, as many ‘free speech’ advocates are pushing for. At some level, there will always be limits, there will always be guardrails, and at times, they may well extend beyond the laws of the land, given the amplification potential of social posts.

There are no easy answers, but leaving it up to the will of the people is not likely to yield a better situation on all fronts.