Teens fed incorrect information by AI ahead of social media reforms
Teens turning to their AI chatbot on Snapchat aren’t getting accurate information about Australia’s upcoming social media reforms, an experiment has shown. Here’s what you need to know
READING LEVEL: ORANGE
Alex Nolastname and Daisy Nolastname are not real teenagers.
But they can still prove real teenagers are being fed incorrect information with few disclaimers* on their favourite social media apps.
The Daily Telegraph recently decided to conduct an experiment. The point of the experiment was to see what teenagers were being told by Snapchat regarding the upcoming social media reforms for under 16s and how those affected would have their accounts archived.
But the chatbot revealed no real information about the new laws, showing that teens who rely on Snapchat’s AI are potentially being left in the dark.
Daily Telegraph reporters created a new profile in which they identified themselves as a 13-year-old boy, Alex Nolastname. They asked their only “friend” in their chat contacts for some advice.
Alex’s AI immediately struck up a friendly tone, greeting him with a chipper “Hi Alex!” and asking, “How can I help today?”.
“R teenagers allowed to use Snapchat,” Alex asked.
The chatbot responded: “Snapchat is for those 13 and older. In some states, guardians of Snapchat users aged 13-17 can make privacy requests for their teens”.
Recognising the app had defaults to US American rules, the reporters clarified: “What about Australia”. It referred Alex to Snapchat Information for French Residents.
“Australia social media ban” the reporters wrote, in a direct bid to find something relevant.
“There’s no nationwide social media ban in Australia,” The AI answered, despite Snapchat being included in a list of platforms that will have a new minimum user age of 16 in Australia from December 10 this year – in less than two weeks’ time.
The misleading answer from AI was concerning, but like any good science experiment, it had to be repeatable to be reliable. The reporters decided to test out the chatbot with another fake account.
Enter “Daisy” – the second reporter-created profile, this time of a 15-year-old girl.
She had a little more luck; her bot acknowledged that there had been “discussion” about social media regulation but insisted there were “no current laws banning them outright”.
Reporter Harry Clifton then also gave it a crack on his own existing account and got similar answers.
In the tech industry, basic bots like Snapchat’s are widely known for being notoriously dumb and outdated.
In its own in-chat pop-up Terms and Conditions, the platform states: “My AI is designed with safety in mind, but may give responses which are biased*, incorrect, harmful or misleading”.
But an increasing number of young people are befriending these bots – a recent US study found as many as one in five teens are spending as much time with an AI companion as their real friends – and the bots are gaining their trust.
Which is a real problem if these AI “friends” aren’t telling the truth.
The experiment shows the importance of turning to trusted adults and reputable* websites to help you navigate the upcoming social reforms, rather than relying on the platforms themselves. The bot experiment is one small example of an impacted platform acting in its own interests by failing to disclose the coming changes – and that kind of omission is probably not in the best interest of any young user.
POLL
GLOSSARY
- disclaimers: a statement that denies something, such as responsibility or liability
- biased: leaning towards a particular viewpoint
- reputable: trustworthy and considered a reliable source of information
EXTRA READING
Massive change coming to Roblox
How U16s can save social data
U16 social media ban explained
QUICK QUIZ
1. What fake profiles of young teens did the reporters use to test Snapchat’s communication with its users?
2. What misinformation did Snapchat’s AI bot feed to the fake teen profiles?
3. Which nation’s rules does Snapchat’s AI bot likely defer to?
4. What proportion of teens are befriending AI chatbots, according to a recent US study?
5. What are basic bots like that used on Snapchat known for within the tech industry?
LISTEN TO THIS STORY
CLASSROOM ACTIVITIES
1. Understand chatbots!
Design a poster or infographic. Your purpose is to help kids understand AI chatbots.
Time: allow at least 30 minutes to complete this activity
Curriculum Links English, Information Technology
2. Extension
What are the rules that need to be followed for scientific experiments? Write a detailed list. Use your research skills to check your ideas. Next to each item on your list, write a sentence explaining why that rule is important.
Time: allow at least 20 minutes to complete this activity
Curriculum Links English, Science, Critical and Creative Thinking
VCOP ACTIVITY
Oral Language / Openers – “Newsreader Challenge”
Time: 5–8 minutes
I can goal: I can read aloud with expression to engage an audience.
Activity:
Choose two short sections from the article.
Students practise reading a section aloud in the style of a TV newsreader — clear pace, expression, and emphasis on key words.
In pairs, one reads while the other listens for expression, clarity, and confidence.
Swap roles.
Invite a few volunteers to be the “Kids News Readers of the Day.”