Written evidence submitted by Marinna Spring (SMH0069)

Social media algorithms have become a bit like tectonic plates, re-shaping our society and how it works. These recommendation systems are baked into the fundamental design of all of the main sites. They aim to maximise engagement, sending users - whether adults or kids - posts and videos they’ll react to. That’s how the companies sell ads and make their money. 

The problem is prioritising engagement over everything else on these sites has consequences. Provocative content - whether true or untrue, harmful and hateful or nice - can be turbocharged and pushed to millions in ways it couldn’t be before. 

From violence and riots to extreme conspiracy theories, misogyny and violent political ideologies, harmful posts pushed to kids and teenagers, the shaping of political opinions and elections - all of these events can be linked back to recommendation systems are reshaping society and the world as we know it. 

I find almost everything I report on now as the BBC’s social media investigations correspondent comes back to this question of the sites algorithms. All of the social media companies say they protect users and ensure their algorithms are safe, and reject this idea. Some - like X - rarely respond to right of reply requests and I have not interviewed a boss at a social media company about any of this for several years. 

Examples of reporting to demonstrate the above:

• For BBC Panorama, Marianna tracked down insiders from the social media companies and teenagers affected by their phones, revealing how social media algorithms show violence to boys. 

Online article: https://www.bbc.co.uk/news/articles/c4gdqzxypdzo

BBC Panorama: https://www.bbc.co.uk/programmes/m0022n0f

• In her BBC Radio 4 podcast series Why Do You Hate Me, Marianna tracked down impersonators, deep-fakers, trolls and conspiracy theorists, as well as those harmed online, to reveal why some people behave the way they do on social media. Lots of the series explored the way that algorithms drive hate and polarisation. In one episode, Marianna investigated how a faked audio clip of London's Mayor talking about pro-Palestinian protests and Armistice Day went viral. In another episode, she interrogated the wave of racist and misogynistic online abuse on X targeting football pundit Eni Aluko. She identified one of the people behind the abusive messages to understand more about why they did it - and he agreed to meet her. 

https://www.bbc.co.uk/programmes/p0jml957/episodes/player
 

https://www.bbc.co.uk/news/uk-68146053
 

• As riots unfolded in the summer, Marianna tracked down the people linked to a pseudo news site accused of fuelling violence with disinformation and revealed their financial motive. This shows how algorithms can be a means for profit. 

https://www.bbc.co.uk/news/articles/c5y38gjp4ygo
 

• For the UK Election, Marianna set up the UK Undercover Voters: 24 fictional people with social media accounts, created to investigate what content is recommended to different types of voters. The profiles are private, with no friends. They just like, follow and watch relevant content. In one of several investigations with these profiles, Marianna revealed how a network of X accounts had been creating and sharing doctored and deepfaked clips of politicians - and then posting misleading comments alongside to bolster the impression they were real. For the first time since Elon Musk took over Twitter, X responded to the BBC’s investigation and removed several of the profiles and posts. This marked real accountability for a social media company during a General Election campaign.

https://www.bbc.co.uk/news/articles/cg33x9jm02ko
 

https://www.bbc.co.uk/news/articles/cq55gd8559eo
 

• Marianna also investigated and analysed how social media impacted the US Election for Americast and the BBC One Election special. For BBC Panorama, she exposed how Donald Trump supporters had been creating AI fakes of black voters to encourage African Americans to vote Republican. She grilled the Trump fans who shared them - often re-purposing satire.

https://www.bbc.co.uk/news/world-us-canada-68440150
 

• For Radio 4’s Why Do You Hate Me USA podcast, Marianna met two women who believed the assassination attempts on Donald Trump were staged – but have very different political views. She interrogated how algorithms and other factors shaped their belief in conspiracy theories and analysed how these have spread across the political spectrum. 

https://www.bbc.co.uk/programmes/p0jql1k6
 

BIO/INFO: Marianna Spring is the BBC’s first social media investigations correspondent. She exposes the real-world harm caused by disinformation, hate and algorithms on social media. Marianna’s journalism has reached millions with impactful investigations and original reporting drawing on her expertise across BBC Panorama, TV bulletins, BBC Breakfast, BBC Radio 4 and podcasts on BBC Sounds and the BBC website. She is emblematic of a new model of multi-media correspondent - now a key part of the BBC’s new Verify team. In 2024, she was awarded both Best Speech Presenter at the ARIAS and Best Host at the British Podcast Awards. Her social media investigations this year have all centred around how algorithms on the main sites prioritise engagement over safety. This has caused real world harm from affecting elections, inflaming riots and impacting teenagers. 

The price of doing this journalism is being a target for online abuse and threats. Marianna is one of the most trolled correspondents at the BBC. She also continues to be threatened outside her place of work and to receive high levels of online abuse. 

26 February 2025