Elon Musk, a self-proclaimed free speech absolutist, says one of the reasons he has taken over Twitter is to fix the platform to strengthen free speech. As he puts it: “Free speech is the bedrock of a functioning democracy, and Twitter is the digital town square where matters vital to the future of humanity are debated.”

He is right. But to strengthen free speech, Musk needs to abandon his ego-centric notion of free speech as “I get to say whatever I want free from censorship” and accept that it is not just about freedom of expression, it is primarily about a ‘free trade of ideas’.

Content moderation, then, is not censorship that impedes free speech, but a necessity to create a space in which free and fair access to speech is possible for everyone, not just the strong and powerful. By keeping harmful speech in check, more, and more diverse ideas will enter the discourse as more people feel safe to speak freely.

But there is a far bigger challenge that Musk must address: the way algorithms used by platforms like Twitter organise and distort this free trade of ideas. Algorithms decide who gets to see what message. They amplify the kind of content that keeps users most ‘engaged’ because engagement means user attention for targeted advertising, and more data collection opportunities – and more money – for the platforms. When I speak on a platform, algorithms effectively determine my audience. Is speech truly free if we can speak, but someone else decides who gets to hear what we have to say?

Algorithmic audience creation amplifies extremist voices and enraging content. The social media platforms have driven polarisation in society because their algorithms have distorted the exchange of ideas. When algorithms decide the reach of each message, speakers who know how to speak to the algorithm can achieve wide circulation of their ideas and engage in large-scale social engineering, while for most, the workings of content distribution algorithms remain opaque.

This needs urgent fixing. It’s a problem more widespread than content moderation because content moderation only affects a minority of harmful speech and algorithmic audience creation applies to all speech on social media. And it is directly linked to the bottom line of social media businesses. The only fix is to change the social media business model that favours targeted advertising.

Encouragingly, Musk has announced his focus is not on making profit, and he wants to strengthen paid subscriptions on Twitter at the expense of relying on advertising. Others have provided ideas for how such a Twitter overhaul could work in an inclusive way, by keeping the majority of accounts free and only charge those that derive value from large followings.

By abandoning algorithmic audience creation and targeted advertising, the free and fair exchange of ideas could be restored. When users hear from those they actually follow and want to hear from, where audiences self-select, public discourse on social media will once again follow democratic principles.