Protecting children from harmful content online has become a top priority for regulators around the world. In the UK, Ofcom, the regulator responsible for enforcing the country’s Online Safety Act, has announced its plans to explore how artificial intelligence (AI) can be used in the fight against harmful content specifically targeted at children.
Ofcom intends to launch a consultation on the use of AI and other automated tools to proactively detect and remove illegal content online. The focus will be on protecting children from harmful content and identifying child sex abuse material that has previously been difficult to detect. This initiative is part of a broader set of proposals aimed at enhancing online child safety.
Mark Bunting, a director in Ofcom’s Online Safety Group, emphasized the importance of accurately assessing the effectiveness of AI tools in identifying and shielding children from harmful content. While some services already utilize these tools, there is limited information available regarding their accuracy and overall effectiveness. Ofcom aims to address this issue, ensuring that industry players assess and manage risks to free expression and privacy when using AI tools.
One potential outcome of this consultation is that Ofcom may recommend how platforms should assess and improve their content blocking capabilities, potentially leading to fines for non-compliance. Bunting stressed that the responsibility for protecting users lies with the platforms themselves, as they must take appropriate steps and utilize suitable tools to ensure online safety.
However, there are both proponents and skeptics of using AI in this context. AI researchers have made significant progress in using AI to detect deepfakes and verify users online. Nevertheless, critics argue that AI detection is not foolproof and may have limitations.
Ofcom’s announcement coincided with the release of its latest research on children’s online engagement in the UK. The study found that an increasing number of younger children are connected to the internet, with 24% of 5- to 7-year-olds owning their own smartphones. Additionally, 38% of children in this age group are already using social media, with Meta’s WhatsApp being the most popular app among them.
The research also revealed that around one-third of children aged 5 to 7 go online independently, with 30% of parents allowing their underage children to have social media profiles. While 76% of parents reported discussing online safety with their young children, there appears to be a disconnect between what children see online and what they report to their parents. Ofcom’s research showed that only 20% of parents were aware of concerning content their children had encountered, despite 32% of the children reporting such experiences.
Furthermore, the study highlighted the challenge of deepfakes, with 25% of children aged 16-17 admitting that they lacked confidence in distinguishing fake from real content online.
Overall, Ofcom’s exploration of AI in combating harmful content is a significant step toward enhancing online child safety. By assessing the effectiveness of AI tools and recommending improvements, regulators can ensure that platforms prioritize protecting children from harmful content. However, it is crucial to acknowledge the limitations of AI detection and continue to explore other strategies to address online safety concerns effectively.