More than a quarter of children consuming news online said they have believed a fake or artificial intelligence-generated story, according to research.

The finding from online safety charity Internet Matters that almost three in 10 (27%) have fallen prey to fake news has prompted calls for children to be given better critical thinking skills to keep them safe online.

The proportion is greater among vulnerable children – which for this report included those who have qualified for special educational needs (SEN) support or have an education, health and care plan (EHCP) or those with a physical or mental health condition.

Around four in 10 (43%) of this group believed a fake or AI-generated story, compared with 23% of children not deemed as vulnerable for this research.

One 17-year-old girl told researchers: “Today I’ve seen about three videos of natural disasters like hurricanes and floods and they’ve all been fake but I believed every single one of them, the AI fooled me.”

The report warned that the spread of misinformation online “can deepen social and political divides and even trigger real-world harm” such as the riots in the aftermath of the Southport murders.

Researchers added: “The spread of AI-generated content, including deepfakes, on social media is also increasing the risk of mis- and disinformation and making it more difficult for users to verify and trust news.”

Deepfakes are defined as AI-generated images, videos and audio manipulated to misrepresent the appearance, speech or actions of someone or something.

Almost half (48%) of children said they feel social media companies should take proactive steps to remove fake news, while 40% said AI-generated content should be clearly flagged or labelled.

Overall, the report found that just over three-quarters (76%) of children consume news weekly, with 68% of those children getting their news from social media.

The proportion getting their information from news outlets’ social media accounts is similar to the proportion who get it from influencers or content creators (41% and 40% respectively).

The charity said social media algorithms play a “significant role” as the research found 40% of children who said they get their news from social media do not actively follow news-focused accounts, with many reporting they come across news content via accounts and posts automatically suggested to them.

The report also warned of the harmful impacts on children if distressing news content is pushed in their online feeds, with 61% of children who get news on social media saying they have seen a story that worries or upsets them in the past month, and 41% reporting feeling overwhelmed by it.

Labour MP Jess Asato, who is a member of the Education Select Committee, said: “Too often, children are exposed to harmful or misleading content online with little support to make sense of it.

“While the Online Safety Act will help to make platforms more accountable, we also need to ensure every child has the skills to navigate the fast-changing digital world safely and critically.

“We need to embed media literacy across education so every child can think critically, question what they see, and stay safe online.”

Rachel Huggins, from Internet Matters, said social media has become “the primary medium of news and information for most children”.

She said: “While social media can offer immediate access to news which keeps children and young people informed and connected with the world around them, the volume of information, which is often negative, poses a risk to their wellbeing.

“The negative effects of this information environment are exacerbated by the rise of mis- and disinformation and AI-generated content, which complicates how to distinguish between true and false online and impacts young people’s trust in the news.

“With the voting age in UK elections potentially set to fall to 16, it is more important than ever that young people are equipped to engage critically with news in an age-appropriate way, particularly the ability to verify information and separate fact from fiction.

“More collective action is needed from the Government, social media platforms, the news sector and schools, to support families on this issue, and enable children and young people to become informed, happy and engaged citizens.”

A Government spokesperson noted that the research pre-dated new safety measures including age checks on sites hosting pornography or other harmful content such as self-harm, suicide or eating disorders.

Those came into force at the end of July under the Online Safety Act but some campaigners and politicians argue there is much more to do to keep children safe on the internet.

Dame Chi Onwurah, Labour MP and chairwoman of the science, innovation and technology committee, said the Act “just isn’t up to scratch to protect users”.

She added: “I urge the Government to take up our recommendations and create a stronger online safety regime that discourages the viral spread of misinformation, regulates generative AI and places much-needed standards onto social media companies.”

A Government spokesperson said as platforms comply with the legal requirements of the Act, this will result in “safer algorithms and less toxic feeds”.

They added: “We’re also working to help families build stronger online safety skills so parents and children can make informed choices online. We will not hesitate to act where evidence shows further intervention is needed to protect children.”

– The report’s findings are based on a survey of 961 UK news-consuming children aged 11-17 in  July 2025.