Spotting Deepfakes: Practical Detection Tips

Deepfakes are AI-made videos and images that show real people doing things they never did. They're getting harder to spot, but a checklist still beats nothing.

9 min · Reviewed 2026

Your eyes will lie to you

Five years ago, deepfakes had giveaway tells: weird hands, wobbling teeth, mouths that didn't match the words. Today, the best ones are almost perfect to a casual eye. So the rule has flipped: you can't really detect a deepfake by looking — you have to investigate the source.

Tells that still sometimes work

Hands and ears: small details still confuse some models
Teeth and earrings: shapes that wobble between frames
Hairlines: smudgy, dissolving edges
Lighting that's wrong on the face vs the background
The voice is a hair too smooth — no normal hesitations or breaths

The source check (more reliable)

Reverse image search: drag the image into Google Images or TinEye
Find the original post and original poster — not a screenshot of a screenshot
Check if a real news organization is reporting the same thing
If the video is shocking, ask: "who would benefit if I believed this?"
Wait 24 hours on outrage. Real news lasts. Hoaxes evaporate.

Trustworthy source signals	Sketchy source signals
Multiple known outlets reporting it	Only a single sus account posted it
Original poster has a long history	Account is brand-new or low-follower
Video has consistent metadata	Watermarks scrubbed off

Try it: detective drill

Find an image that went viral this week. Reverse-image-search it. Trace it back to the first place it was posted. Check whether real outlets are reporting on it. The whole process takes about 5 minutes once you've done it once.

End-of-lesson check

8 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-builders-spotting-deepfakes-builders

What is the main idea of "Spotting Deepfakes: Practical Detection Tips"?
1. Deepfakes are AI-made videos and images that show real people doing things they never did.
2. Use AI as the final authority for the whole decision
3. Avoid checking the answer once it sounds polished
4. Focus only on speed instead of judgment
Which concept is most central to "Spotting Deepfakes: Practical Detection Tips"?
1. media literacy
2. deepfake
3. verification
4. synthetic media
Which use of AI fits this topic best?
1. Let the AI decide what matters without your review
2. Use the answer before checking whether it fits the situation
3. Hands and ears: small details still confuse some models
4. Use the first answer without checking it
What should a careful learner remember about "These tells are disappearing"?
1. Each new model fixes the last batch of giveaways. Don't trust visual checks alone.
2. Skip the context so the tool can guess faster
3. Treat the output as private even after sharing it online
4. Use the answer without checking the source
You want to use AI after this lesson. What is the safest next step?
1. Act immediately because the AI answer is written clearly
2. AI cannot make the human values or safety decision for you.
3. Hide uncertainty so the final answer looks cleaner
4. Use private or sensitive details before checking permission
How should AI output about deepfake be treated?
1. As proof that no other source is needed
2. As a replacement for context, consent, or expert review
3. As a draft or helper output that still needs human judgment and verification
4. As something that becomes correct when it sounds confident
Name one way to verify an AI answer about deepfake.
Which action would help you apply "Spotting Deepfakes: Practical Detection Tips" responsibly?
1. Use the tool to avoid thinking through the tradeoff
2. Keep going even if the output conflicts with a trusted source
3. Use the first answer without checking it
4. Teeth and earrings: shapes that wobble between frames

← Back to interactive lesson

Tendril · Builders · Safety & Governance

Spotting Deepfakes: Practical Detection Tips

Deepfakes are AI-made videos and images that show real people doing things they never did. They're getting harder to spot, but a checklist still beats nothing.

9 min · Reviewed 2026

Your eyes will lie to you

Tells that still sometimes work

Hands and ears: small details still confuse some models
Teeth and earrings: shapes that wobble between frames
Hairlines: smudgy, dissolving edges
Lighting that's wrong on the face vs the background
The voice is a hair too smooth — no normal hesitations or breaths

The source check (more reliable)

Reverse image search: drag the image into Google Images or TinEye
Find the original post and original poster — not a screenshot of a screenshot
Check if a real news organization is reporting the same thing
If the video is shocking, ask: "who would benefit if I believed this?"
Wait 24 hours on outrage. Real news lasts. Hoaxes evaporate.

Trustworthy source signals	Sketchy source signals
Multiple known outlets reporting it	Only a single sus account posted it
Original poster has a long history	Account is brand-new or low-follower
Video has consistent metadata	Watermarks scrubbed off

Try it: detective drill

End-of-lesson check

8 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-builders-spotting-deepfakes-builders

What is the main idea of "Spotting Deepfakes: Practical Detection Tips"?
1. Deepfakes are AI-made videos and images that show real people doing things they never did.
2. Use AI as the final authority for the whole decision
3. Avoid checking the answer once it sounds polished
4. Focus only on speed instead of judgment
Which concept is most central to "Spotting Deepfakes: Practical Detection Tips"?
1. media literacy
2. deepfake
3. verification
4. synthetic media
Which use of AI fits this topic best?
1. Let the AI decide what matters without your review
2. Use the answer before checking whether it fits the situation
3. Hands and ears: small details still confuse some models
4. Use the first answer without checking it
What should a careful learner remember about "These tells are disappearing"?
1. Each new model fixes the last batch of giveaways. Don't trust visual checks alone.
2. Skip the context so the tool can guess faster
3. Treat the output as private even after sharing it online
4. Use the answer without checking the source
You want to use AI after this lesson. What is the safest next step?
1. Act immediately because the AI answer is written clearly
2. AI cannot make the human values or safety decision for you.
3. Hide uncertainty so the final answer looks cleaner
4. Use private or sensitive details before checking permission
How should AI output about deepfake be treated?
1. As proof that no other source is needed
2. As a replacement for context, consent, or expert review
3. As a draft or helper output that still needs human judgment and verification
4. As something that becomes correct when it sounds confident
Name one way to verify an AI answer about deepfake.
Which action would help you apply "Spotting Deepfakes: Practical Detection Tips" responsibly?
1. Use the tool to avoid thinking through the tradeoff
2. Keep going even if the output conflicts with a trusted source
3. Use the first answer without checking it
4. Teeth and earrings: shapes that wobble between frames

← Back to interactive lesson