Helping children navigate online misinformation in the age of generative AI

We all agree that children need protection from a host of dangers in the physical world. 难道在数字世界中不应该是一样的吗? The reality is, online platforms are playing a larger role in children’s lives than ever before. 加速 大流行, this heightened usage goes beyond merely facilitating school education and social networking.  

Children develop pivotal social and political opinions during their formative years, 其中大部分现在都是在网上形成的. A 报告 by the National Literacy Trust found that more than half of 12-15 year olds regularly use social media as their main news source. 然而，他们升高了 脆弱性 这也使他们成为错误信息传播者的主要目标.

它比以往任何时候都更重要, 然后, that we examine the factors that continue to drive online misinformation and implement systems that protect children from its negative impact.
 

生成式人工智能是如何放大网络错误信息的

专家 项目 that nearly 90 per cent of online content may be generated or 操作d with AI by 2026. While much of this will be using AI to push content to even greater heights, it also makes the creation of misleading content and websites easier. 考虑到这一点，这尤其令人担忧 11% of school children can distinguish a fake website from a real one.

Deepfakes, 例如, are generative AI videos and images that convincingly replicate the mannerisms and speech of real people — or create digital personas from scratch. These are already being used to distort information on social and political issues, and 操作 为了经济利益而生孩子. 深度造假已经出现 排名 as the most serious AI crime threat due to their uncanny ability to deceive people and their increasingly easy production, 复制, 和可访问性. Given children’s lack of familiarity with such 操作d content, deepfakes are also 预期 对他们构成更大的威胁.

更重要的是，人工智能产生的错误信息已经被 所示 甚至比人类创造的虚假内容更有说服力. 在某些情况下，人工智能生成的人脸是 认为 比真人更值得信赖. In its active crackdown of fake profiles and content, more than 三分之二 of the accounts that Meta removed from its platforms in 2022 likely had AI-generated profile pictures.
 

如何让网络空间对儿童更安全

因为生成式人工智能可能会增加网上的错误信息, 法规的组合, 教育系统的实际变化, and technological interventions must be implemented to keep children safe.

政策makers should formulate standardised procedures that technology companies can use to classify the safety and reliability of content across their platforms, 确保用户的透明度和问责制, 广告商, 和出版商. Industry-wide 标准, such as those laid out by the Global Alliance for Responsible Media (加姆), also act to align 广告商 和出版商 on content safety classifications. These 标准 can help inform brands on where to direct their advertising budgets, guaranteeing appearing next to and the funding of content that doesn’t harm young people — and inadvertently, 这个品牌.

除了, children require up-to-date education on how to spot misleading or AI-generated posts online. This includes being taught how to look for inconsistencies in fake content, 例如奇怪的短语或不自然的视觉组件, 与有信誉的新闻来源交叉核对网上的内容, 核实服务, 或者是更了解事实信息的成年人. 

而科技平台可以发挥作用, brands can help stifle misinformation by ensuring their ad spend doesn’t fund or endorse distributors of misinformation. AI-powered brand safety solutions are increasingly effective at identifying online misinformation. By comprehensively analysing the contextual environment and intent of dynamic forms of media, these solutions provide an easy way for marketers to understand content safety — which allows them to fine-tune their media plans accordingly. 由于人工智能的检测能力仍在发展中, the accompanying nuances of online content can be more thoroughly understood and identified when human moderators also work in combination with AI.

The internet is a remarkable resource for children as it opens the door to a diverse range of educational material, 加强社会关系, and provides opportunities for seeking support where it may not be possible in immediate environments — we should not stop young people from using it. 

The focus must be on cultivating safe and positive online environments where children can thrive and reap the benefits of the internet. This is best achieved through implementing progressive policies and innovative AI-powered solutions that ensure the safety of content online, while updating the education system to equip children with a robust educational foundation for avoiding the pitfalls of generative AI and online misinformation.