Every week, Talking Points gives you a worksheet to practise your reading comprehension with exercises about the story we’ve written.
The internet is drowning in “AI slop” – low-quality content generated by artificial intelligence that makes it difficult to tell what is real. Children are particularly vulnerable to this uncensored flood of videos, memes and more.
Hong Kong educational psychologist Dr Kitty Yeung Kit-yu expressed concern over the lack of regulation of AI videos, the inappropriate content in them as well as their impact on development and learning.
While the internet has long been chipping away at protections for children, AI’s ease in mass-producing content combined with social media’s addictive algorithms has exacerbated this problem – making it harder for parental supervision to keep up.
No regulation
Yeung said that unlike traditional films or shows with age classifications, online videos operate without any guardrails.
“Sometimes I see short clips or even hour-long videos that just pop up without you needing to register for anything, and I feel the content is essentially [for those aged] 18 plus. They might not show highly graphic visuals, but the text and dialogue are clearly [adult content],” Yeung said.
She pointed out that parents might not be suspicious of colourful animations or well-known cartoon characters and would not examine them more closely for mature themes.
“Even with animated characters, there are sexual elements or … suggestions,” Yeung said. “I believe these AI-generated videos aren’t meant for children, but they are produced in a cartoon format. If children … watch them without questioning the content, I think it’s dangerous.”
Harmful tropes
Online algorithms could also amplify clickbait narratives popular in AI content, Yeung said. Popular tropes include impoverished characters instantly becoming billionaires, toxic revenge plots or storylines framing marriage to a wealthy partner as life’s ultimate goal.
She warned that these stories promote greed, normalise toxic relationships and blur moral boundaries regarding how people should treat one another.
“A common plot I see is a girl from a very poor family suddenly discovering she is the daughter of a billionaire. Or a girl and boy get together, and she finds out he is incredibly rich but was deceiving her,” the psychologist explained.
She expressed concern about children learning their world view from these AI-generated videos.
“When watching these, will children get the wrong idea about life and role models? Will they start thinking, ‘If I have money, I can do whatever I want,’ or ‘My goal is just to get rich so I can step on others’?” Yeung asked.
“It’s not that they can’t watch, but we have to teach them the context: these are just movies portraying abnormal lifestyles that don’t represent how most people actually behave.”
She added: “Because of this, I think there needs to be regulation, or at the very least, parents need to pay close attention.”
Impact on learning
The harm could go deeper for young children, Yeung said. Toddlers aged two to three who rely on clear input for language development are being exposed to unnatural, robotic AI voices featuring odd pitches and fragmented grammar, the psychologist noted.
Children might enjoy the visuals, but Yeung warned that this continuous exposure to robotic cadence could compromise foundational speech skills.
“Kids constantly hear sentences that are awkwardly chopped up, with unnatural breathing pauses, weird pitches or odd phrasing,” she said. “Are the phrases, grammar and vocal tones they are picking up actually appropriate?”
Worse still are videos featuring artificial gibberish: “A very young child might actually think it is a real language and try to imitate it when learning to speak,” Yeung said.
For older children, it can be hard for them to distinguish if content has been doctored or twisted by AI.
Sixteen-year-old Angela Qian expressed concern over AI’s ability to manipulate information.
“Users with malicious intent can now fabricate convincing videos from a single text prompt,” said Angela, a Year 12 student at German Swiss International School. “Because these visual formats are inherently more engaging than plain text … viewers are far more likely to believe them.”
She added that AI-generated media was becoming increasingly difficult to spot.
“Individuals with less digital experience simply lack the tools to distinguish real content from synthetic fakes,” Angela said.
Advice for parents
Yeung urged parents to take an active role in curating and vetting the content their children consume.
“Don’t just hand them a screen because they won’t sit still during meals. If you really feel you must give them a screen, you can control the content,” she said.
Rather than leaving children to scroll through short videos, she recommended directing them to established platforms with built-in quality controls.
“If we really want to let children watch something, I genuinely think platforms like Netflix or Disney have many shows that are much more suitable,” the psychologist said.
“They can watch those instead of these short, low-quality clips, which I find harmful because they lack quality control.”
