Social media platforms don’t work like they used to. When Facebook first introduced its News Feed feature in 2006, the feed showed us what our friends were doing and allowed us to follow news outlets, celebrities, or other pages of our choosing.  

Now, algorithms are in great part responsible for what you see online. They have a huge impact on how people perceive digital spaces and what kind of content, or dangerous people, they interact with.  

Despite their importance in people’s daily lives, social media algorithms are often not designed with your safety in mind. They are designed to push highly engaging and addictive content to users to increase engagement and commercial transactions. This engaging content is often upsetting or harmful and can include violent threats, antisemitism, content encouraging self-harm and eating disorders, illicit drug sales, and other forms of harm. 

Social media companies design their algorithms in this way to keep people scrolling on their platforms for as long as possible. Because that’s how they make their money, despite people’s mental health and wellbeing. Engagement is now about profit, not the public good.

Keep reading to learn: 

• What is an algorithm? 
• How do social media algorithms work? 
• How are algorithms shaping the way we think? 
• Why do social media algorithms enable and amplify harmful content? 
• What can social media companies and policymakers do to make algorithms and platforms safer? 

What is an algorithm?

An algorithm is a set of step-by-step instructions for solving a problem or completing a task. It tells a computer exactly what to do in order to reach a specific outcome. 

On social media, algorithms set the rules for how the platform will operate – including ranking content, recommending a video, filtering out hate speech or promoting it, and even shadow banning an account (when users have their content visibility limited by the platform).  

Algorithms have a huge impact on what kind of content users interact with. It’s what makes you see what you see. 

How do social media algorithms work?

Social media algorithms work as a filtering system, selecting which posts will get priority for different audiences.  

These algorithms are programed to boost user engagement. This means they get your attention and keep you scrolling on the platform for as long as possible. They largely drive content based on what the platform knows about you, your preferences, and your online behavior. 

Here’s an example of how it works: If you like a post or watch a few seconds of a video about gardening, the algorithm will likely recommend more content on this subject and can even recommend garden centers near you if the platform has your location information. And say you follow a football player, the system will likely suggest that you follow other players and also show you sports content. 

All of that sounds harmless enough, unfortunately these algorithm-driven recommendation systems are often anything but. 

Social media algorithms are too often designed without your safety in mind. They’re at the heart of platform design – yet, there is clear evidence they cause real world harm. But in the U.S., you can’t do much about it because of Section 230. 

Section 230 is a law written in 1996 to support the growth of the early internet, long before social media, when online platforms were largely message boards and had no algorithms, no personalized feeds, and no meaningful ability to shape what people saw. Lawmakers wanted to encourage companies to remove harmful material without fear of being sued, so they gave them broad legal protections. The internet has changed dramatically since then, but the law has not. Today, it means companies cannot be sued, even when their algorithms amplify harmful content that reaches children or spreads dangerous material.  

Every other major industry in America can be held accountable when their products cause harm. Think automobile, food, and drug safety standards. But tech platforms are the exception – they receive special treatment from an outdated law. 

How are algorithms shaping the way we think?

When Elon Musk took over Twitter in 2022, and renamed it X, he said the platform would be “the digital town square where matters vital to the future of humanity are debated.” But the way things are currently allowed to operate shows that the idea that social media platforms are venues for fair public debates is not just a stretch, it’s a fallacy. And that’s because algorithms are acting in the background to promote certain voices and ideas. 

These don’t work just to personalize your feed. They also push popular content that is getting traction on the platform. Controversial content like conspiracy theories, misinformation, hate speech or other online harms often go viral because they generate emotional responses – whether people are endorsing or disagreeing with it. Responses of any kind drive profit for these companies. 

When U.S conservative icon Charlie Kirk was shot in September 2025, X was rapidly flooded not only by the horrific video of the attack, but also by hate and false narratives. CCDH’s research showed that X posts celebrating Kirk’s death were viewed 52 million times, while posts calling for retaliatory violence were seen 43 million times – even though they represented only 1.2% and 0.4% respectively of posts about Kirk.  

The amplification of controversial content about Kirk’s murder created the impression that most people on X were glorifying or inciting violence – but that was a trick of the algorithm, as our research shows. 

Something similar happened after the Southport stabbings in the UK, when a man murdered three young girls at a dance class in 2024. The tragedy was followed by a wave of hate and misinformation on social media, even before the attacker had been identified. At the time, our researchers found that X was enabling and profiting from hateful and misleading posts about Muslims and migrants being responsible for the incident. These lies directly fueled violent summer riots in the UK. 

Why do social media algorithms enable and amplify harmful content?

That’s easy, to make money. Social media companies design their algorithms to keep people hooked on their platforms, which harmful content does. It lets the companies accumulate data on users and then sell it to advertisers. 

These companies make billions of dollars on the back of harmful content. Read our explainer to learn more about what advertisers can do to help stop the problem. 

Because of Section 230 in the U.S., and the absence of a regulatory agency to oversee social media, companies have so little incentive to curb the spread of online harms that some have dismantled their content moderation teams and slashed safety policies, allowing the algorithms to run largely unencumbered.  

Leaked internal guidance, for example, showed that Meta now permits racist statements like “Immigrants are grubby, filthy pieces of sh*” and “Jews are flat out greedier than Christians” on their platforms. 

What can social media companies and policymakers do to make algorithms and platforms safer?

The first and most important action social media companies should take is to be transparent about how their algorithms work. Transparency is essential for policymakers and society to have a clear understanding of these companies’ priorities and the potential harms they can cause. 

Tech leaders should:   

• Be transparent about how their algorithms select and rank content on the platform. 
• Institute safety by design, considering people’s safety from the moment the algorithm is being designed. 
• Institute strong enforcement policies that target scammers, terrorists, predators, harassers, stalkers, and other criminal actors.  

Governments and lawmakers should:   

• In the U.S., reform Section 230 to clarify that algorithmic recommendations are outside the scope of the law’s immunity shield.  
• Create and enforce regulations and duty of care requirements to hold social media platforms accountable and responsible for what is amplified by their platforms’ algorithms, including for the harms caused to children. 
• Implement the four principles of CCDH’s STAR Framework for social media regulation: Safety-by-design, Transparency, Accountability, and Responsibility.    

---