How to spot AI-generated images
Look for strange hands, warped text, mismatched reflections, odd shadows, and details that do not follow real-world logic.
Read More
Resources
Use these beginner-friendly resources as workshop handouts, social media post ideas, or starting points for school discussions.
Quick learning cards
Look for strange hands, warped text, mismatched reflections, odd shadows, and details that do not follow real-world logic.
Read MoreCompare sources, check dates, look for original evidence, and be careful with posts designed to create anger or panic.
Read MoreDeepfakes are AI-edited media that can make someone appear to say or do something they did not actually say or do.
Read MoreAI tools can sometimes give confident answers that are wrong, made up, outdated, or missing important context.
Read MoreUse AI to brainstorm and learn, but check facts, protect private information, and follow school rules on academic honesty.
Read MoreAI systems can reflect unfair patterns in data, so outputs should be checked carefully and not treated as automatically neutral.
Read MoreRead more
Start with the source, then inspect the image. Warning signs include unnatural hands, unreadable background text, inconsistent lighting, repeated patterns, distorted objects, or faces that look smooth but slightly unrealistic.
Do not rely on one screenshot or one account. Search the claim, compare reliable sources, check when it was published, and ask whether the post is trying to make you react quickly before thinking.
A deepfake can imitate a person's face, voice, or movements. Treat surprising clips with caution, especially if they ask for money, private information, or urgent action.
An AI hallucination is when an AI tool presents incorrect information as if it is true. Always verify important facts using trusted sources, especially for schoolwork, health, money, or legal topics.
Good student use means using AI as a learning helper, not as a replacement for thinking. Avoid sharing private information, cite help when required, and check answers before submitting work.
AI tools learn from data made by people and systems. If that data contains unfair patterns, the AI output may repeat them. This is why people should question results and include human judgement.
Local case studies
These placeholders can later become full posts or workshop slides after the team checks dates, sources, and screenshots.
Placeholder case study: posts falsely claimed Woodlands MRT was closed for disinfection during the early COVID-19 period. A useful lesson is to check official transport and government updates before sharing service-disruption claims.
SourcePlaceholder case study: Singapore Police warned about scams using deepfake video to impersonate senior government officials. A useful lesson is that realistic video meetings still need independent verification.
Source