Settings
Language
Color Theme
What Are Deepfakes?
Deepfakes are synthetic media, usually videos, images, or audio, created using artificial intelligence to convincingly mimic real people. They can make someone appear to say or do something they never actually did.
Which of the following BEST describes what a deepfake is?
Deepfakes are AI-generated videos or audio that mimic real people. They’re designed to look and sound convincing, often used to spread misinformation or commit fraud by faking someone’s voice or face. That’s what makes them dangerous, and effective.
View Options Again
A fake website used for phishing attack
A digitally altered image used for social media filters
A secure backup of your personal data
AI-generated video or audio designed to mimic real people
Core Practices for Spotting Deepfakes
Protecting yourself against malicious deepfakes requires you to understand how deepfakes are created, how they’re used, and how you can spot the subtle signs of video or audio manipulation. Over the coming pages, we’ll explore these concepts further.
Recognizing Manipulated Videos
Recognizing manipulated videos is a critical part of deepfake awareness. Even the most convincing videos can contain subtle flaws that reveal they’ve been altered. This includes unusual facial movements, mismatched lighting, unnatural speech patterns, or behaviors that seem out of character for the person.
Which of the following is a possible sign that a video may be a deepfake?
Deepfake videos often struggle to perfectly match lip movements with speech, creating a slight delay or mismatch.
View Options Again
Smooth, natural speech with emotional tone
Out-of-sync voice and mouth movements
Lighting and shadows that match the environment perfectly
Perfectly natural facial expressions
Question Authenticity
Never take visual or audio content at face value, especially when the stakes are high. Treat unexpected, urgent, or sensitive requests with caution and investigate further before responding.
Verify Identities
Always confirm the person’s identity using known secure communication channels. Call them directly, arrange a face-to-face meeting, or use a previously verified contact method before acting on any request.
Which is the safest response to an unexpected request for sensitive information on a video call?
Deepfakes and impersonation scams can make video calls appear genuine. Visual and audio cues aren’t always enough to confirm someone's identity. If the stakes are high, verify someone’s identity through a secure, trusted communication channel.
View Options Again
Record the call for later review
Verify the request using a trusted communication channel
Share the information immediately
Ask them security questions you just made up on the spot
Which of the following is NOT a key component of deepfake awareness?
Deepfake awareness is about protecting yourself from manipulation, not creating it. The focus is on spotting fakes, questioning what you see or hear, verifying authenticity, and staying informed.
View Options Again
Learning to create realistic deepfakes
Verifying identities
Recognizing manipulation
Questioning authenticity
Is the following statement True or False:
Seeing or hearing someone in a video call is always enough to confirm their identity.
Deepfakes can mimic a person’s face and voice in real time, making sight and sound alone unreliable for identity verification.
View Options Again
True
False
Is the following statement True or False:
Deepfake technology is only a risk to celebrities and public figures.
Scammers increasingly target everyday individuals, colleagues, and family members using deepfakes to exploit trust.
View Options Again
True
False
Why Deepfake Awareness Matters
Deepfake scams can lead to financial loss, reputational damage, data breaches, and erosion of trust. By staying alert and applying deepfake awareness practices, individuals and organizations can prevent costly mistakes and protect their integrity.
Which of the following is a potential consequence of failing to spot a deepfake scam?
Falling for a deepfake scam can have serious consequences, including money loss, stolen information, and damaged trust.
View Options Again
More efficient workflow and reduced security checks
Financial loss, reputational harm, and data compromise
Improved AI training data
Increased trust in digital communications
Eye Movement
Eyes are hard to fake. In real footage, blinks are irregular, gaze shifts, and pupils react to light. Deepfakes miss these cues and may have missing or mechanical blinks, fixed stares, over-bright eyes or static pupils. Take note of the deepfake video below, where the speaker doesn't blink.
Which eye behavior is a strong red flag for a deepfake?
Humans blink irregularly, especially during conversation. Deepfakes often suppress or mistime blinks because generators prioritize face/voice alignment and struggle with involuntary micro-movements. The result is long, fixed stares or robotic blink bursts.
View Options Again
Irregular, natural blinking
Pupils constrict when light increases
No blinking for extended periods
Brief glances off-camera
Lip Sync Mismatch
Deepfakes tend to struggle when it comes to lip-syncing. Real speech syncs lip, jaw, and cheek movement with audio. Deepfakes often have a slight lag, wrong mouth shapes, or mouth movement that doesn’t line up.
Which sign suggests a video may be a deepfake?
Real speech has tight timing between sounds and mouth shapes. Deepfakes often miss that micro-timing, so lips lag the audio or form the wrong shapes.
View Options Again
Natural head movement and facial expressions
Out-of-sync voice and mouth movements
Consistent lighting across the scene
Smooth camera focus
Background & Environmental Mismatches
Audio–visual mismatch is a deepfake tell. If the voice sounds studio-clean while the person is outdoors, if echo appears without a source, or clothes ripple with silent wind, something’s off.
What’s the most telling sign that this video is an AI-generated deepfake?
Deepfakes often struggle to replicate subtle human expressions. While a face might blink or smile, it can still look off, with expressions that are stiff, delayed, or robotic. This lack of nuance is a strong giveaway. Real human emotion involves micro-expressions and muscle movements that are hard for AI to fake convincingly.
View Options Again
The subject’s facial expressions seem rigid or unnatural
The mouth movements are perfectly in sync with the voice
The lighting is consistent across the scene
A passing bus briefly blocks the subject
The Importance of Deepfake Awareness
By recognising manipulation, questioning authenticity, verifying identities, and staying informed, you can reduce the risk of falling victim to deepfake scams. Awareness today means protection tomorrow, for you and your organization.
Back
Next
Translation Correction
×
Select a translation to correct...
Begin the Guided Tour
[0, "Begin the Guided Tour", "Begin the Guided Tour"],
Look out for urgency in email subjects, fraudulent sender addresses or requests to perform an action.
[1, "Look out for urgency in email subjects, fraudulent sender addresses or requests to perform an action.", "Look out for urgency in email subjects, fraudulent sender addresses or requests to perform an action."],
Compose
[2, "Compose", "Compose"],
Folders
[3, "Folders", "Folders"],
Inbox
[4, "Inbox", "Inbox"],
Starred
[5, "Starred", "Starred"],
Draft
[6, "Draft", "Draft"],
Sent Mail
[7, "Sent Mail", "Sent Mail"],
Spam
[8, "Spam", "Spam"],
Trash
[9, "Trash", "Trash"],
[URGENT] Claim Your Work From Home Set-Up Payment
[10, "[URGENT] Claim Your Work From Home Set-Up Payment", "[URGENT] Claim Your Work From Home Set-Up Payment"],
Human Resources
[11, "Human Resources", "Human Resources"],
to
[12, "to", "to"],
Claim Your Work From Home Set-Up Payment
[13, "Claim Your Work From Home Set-Up Payment", "Claim Your Work From Home Set-Up Payment"],
Our company acknowledges that we are all working from home for a longer period and is offering a one-time payment to all employees of
[14, "Our company acknowledges that we are all working from home for a longer period and is offering a one-time payment to all employees of", "Our company acknowledges that we are all working from home for a longer period and is offering a one-time payment to all employees of"],
to ensure that you have a suitable home working set-up.
[15, "to ensure that you have a suitable home working set-up.", "to ensure that you have a suitable home working set-up."],
To receive the payment via payroll
[16, "To receive the payment via payroll", "To receive the payment via payroll"],
you will need to complete this
[17, "you will need to complete this", "you will need to complete this"],
acknowledgement form
[18, "acknowledgement form", "acknowledgement form"],
For more information on how to set up your home office space safely, please look at the
[19, "For more information on how to set up your home office space safely, please look at the", "For more information on how to set up your home office space safely, please look at the"],
Thank you
[20, "Thank you", "Thank you"],
The Human Resources Team
[21, "The Human Resources Team", "The Human Resources Team"],
This is an automatically generated email, please do not reply
[22, "This is an automatically generated email, please do not reply", "This is an automatically generated email, please do not reply"],
Is the following statement True or False
[23, "Is the following statement True or False", "Is the following statement True or False"],
View Options Again
[24, "View Options Again", "View Options Again"],
Email Subject: Urgent Action
[25, "Email Subject: Urgent Action", "Email Subject: Urgent Action"],
Phishing attacks are designed to put a
[26, "Phishing attacks are designed to put a", "Phishing attacks are designed to put a"],
time pressure on us to act fast
[27, "time pressure on us to act fast", "time pressure on us to act fast"],
This can cause us to skip much of the critical thinking we normally apply when browsing our emails
[28, "This can cause us to skip much of the critical thinking we normally apply when browsing our emails", "This can cause us to skip much of the critical thinking we normally apply when browsing our emails"],
Email Sender: Fradulent Address
[29, "Email Sender: Fradulent Address", "Email Sender: Fradulent Address"],
Email Sender: Fraudulent Address
[30, "Email Sender: Fraudulent Address", "Email Sender: Fraudulent Address"],
Attackers will often use
[31, "Attackers will often use", "Attackers will often use"],
obscure email addresses and use display names
[32, "obscure email addresses and use display names", "obscure email addresses and use display names"],
that appear legitimate to the naked eye. Be cautious and carefully inspect email sender information
[33, "that appear legitimate to the naked eye. Be cautious and carefully inspect email sender information", "that appear legitimate to the naked eye. Be cautious and carefully inspect email sender information"],
Email Content: Engaging Topic
[34, "Email Content: Engaging Topic", "Email Content: Engaging Topic"],
Attackers often use a
[35, "Attackers often use a", "Attackers often use a"],
broad but important topic
[36, "broad but important topic", "broad but important topic"],
to increase the likelihood of a victim interacting with the phishing material. These topics may include geographic, political or financial themes.
[37, "to increase the likelihood of a victim interacting with the phishing material. These topics may include geographic, political or financial themes.", "to increase the likelihood of a victim interacting with the phishing material. These topics may include geographic, political or financial themes."],
Email Link: Phishing Website
[38, "Email Link: Phishing Website", "Email Link: Phishing Website"],
hovering over the link
[39, "hovering over the link", "hovering over the link"],
you'll see the true link location. Often this is enough to see the malicious intent.
[40, "you'll see the true link location. Often this is enough to see the malicious intent.", "you'll see the true link location. Often this is enough to see the malicious intent."],
Wrapping up
[41, "Wrapping up", "Wrapping up"],
If you
[42, "If you", "If you"],
spot anything suspicious
[43, "spot anything suspicious", "spot anything suspicious"],
with the email sender, subject, content, links or attachments
[44, "with the email sender, subject, content, links or attachments", "with the email sender, subject, content, links or attachments"],
Don't take the risk. Report the email to your IT or Security team for review.
[45, "Don't take the risk. Report the email to your IT or Security team for review.", "Don't take the risk. Report the email to your IT or Security team for review."],
Submit
[46, "Submit", "Submit"],
Next
[47, "Next", "Next"],
Back
[48, "Back", "Back"],
Malicious links will often appear with innocent looking text. By
[49, "Malicious links will often appear with innocent looking text. By", "Malicious links will often appear with innocent looking text. By"],
Correct!
[50, "Correct!", "Correct!"],
Incorrect
[51, "Incorrect", "Incorrect"],
True
[52, "True", "True"],
False
[53, "False", "False"],
Change language
[54, "Change language", "Change language"],
Current Language
[55, "Current Language", "Current Language"],
Use Browser Settings
[56, "Use Browser Settings", "Use Browser Settings"],
Select a language
[57, "Select a language", "Select a language"],
Change Language
[58, "Change Language", "Change Language"],
Please select a language
[59, "Please select a language", "Please select a language"],
Language
[60, "Language", "Language"],
Color Theme
[61, "Color Theme", "Color Theme"],
Color Theme And Background Selector
[62, "Color Theme And Background Selector", "Color Theme And Background Selector"],
Colorful Theme
[63, "Colorful Theme", "Colorful Theme"],
Select Theme
[64, "Select Theme", "Select Theme"],
Close
[65, "Close", "Close"],
Dark Theme
[66, "Dark Theme", "Dark Theme"],
Green Theme
[67, "Green Theme", "Green Theme"],
Purple Theme
[68, "Purple Theme", "Purple Theme"],
Light Theme
[69, "Light Theme", "Light Theme"],
Blue Theme
[70, "Blue Theme", "Blue Theme"],
Settings
[71, "Settings", "Settings"],
Theme Selected
[72, "Theme Selected", "Theme Selected"],
This training has been translated from English. If there is an inaccuracy, please report the correct translation by clicking here!
[73, "This training has been translated from English. If there is an inaccuracy, please report the correct translation by clicking here!", "This training has been translated from English. If there is an inaccuracy, please report the correct translation by clicking here!"],
Translation Correction
[74, "Translation Correction", "Translation Correction"],
Select a translation to correct...
[75, "Select a translation to correct...", "Select a translation to correct..."],
Original Text
[76, "Original Text", "Original Text"],
Correct Translation
[77, "Correct Translation", "Correct Translation"],
Translation to correct...
[78, "Translation to correct...", "Translation to correct..."],
Submissions are reviewed and applied within 48 hours.
[79, "Submissions are reviewed and applied within 48 hours.", "Submissions are reviewed and applied within 48 hours."],
Submit Correction
[80, "Submit Correction", "Submit Correction"],
Submission Successfully Sent
[81, "Submission Successfully Sent", "Submission Successfully Sent"],
Submission Error - Please Try Again
[82, "Submission Error - Please Try Again", "Submission Error - Please Try Again"],
Submission Sending...
[83, "Submission Sending...", "Submission Sending..."],
Mandatory Viewing
[84, "Mandatory Viewing", "Mandatory Viewing"],
Please watch the video from beginning to end before proceeding.
[85, "Please watch the video from beginning to end before proceeding.", "Please watch the video from beginning to end before proceeding."],
Ok
[86, "Ok", "Ok"],
Note: The Colorful, Blue, Light, and Dark Themes are all WCAG 2.2 Level AA conformant.
[87, "Note: The Colorful, Blue, Light, and Dark Themes are all WCAG 2.2 Level AA conformant.", "Note: The Colorful, Blue, Light, and Dark Themes are all WCAG 2.2 Level AA conformant."],
Language translation for this training module has been disabled and explicitly set to the following language
[88, "Language translation for this training module has been disabled and explicitly set to the following language", "Language translation for this training module has been disabled and explicitly set to the following language"],
Deepfake-Awareness
[89, "Deepfake-Awareness", "Deepfake-Awareness"],
What Are Deepfakes?
[90, "What Are Deepfakes?", "What Are Deepfakes?"],
Deepfakes are synthetic media, usually videos, images, or audio, created using artificial intelligence to convincingly mimic real people. They can make someone appear to say or do something they never actually did.
[91, " Deepfakes are synthetic media, usually videos, images, or audio, created using artificial intelligence to convincingly mimic real people. They can make someone appear to say or do something they never actually did.", " Deepfakes are synthetic media, usually videos, images, or audio, created using artificial intelligence to convincingly mimic real people. They can make someone appear to say or do something they never actually did."],
Which of the following BEST describes what a deepfake is?
[92, "Which of the following BEST describes what a deepfake is? ", "Which of the following BEST describes what a deepfake is? "],
A secure backup of your personal data
[93, "A secure backup of your personal data", "A secure backup of your personal data"],
A digitally altered image used for social media filters
[94, "A digitally altered image used for social media filters", "A digitally altered image used for social media filters"],
AI-generated video or audio designed to mimic real people
[95, "AI-generated video or audio designed to mimic real people", "AI-generated video or audio designed to mimic real people"],
A fake website used for phishing attack
[96, "A fake website used for phishing attack", "A fake website used for phishing attack"],
Deepfakes are AI-generated videos or audio that mimic real people. They’re designed to look and sound convincing, often used to spread misinformation or commit fraud by faking someone’s voice or face. That’s what makes them dangerous, and effective.
[97, "Deepfakes are AI-generated videos or audio that mimic real people. They’re designed to look and sound convincing, often used to spread misinformation or commit fraud by faking someone’s voice or face. That’s what makes them dangerous, and effective. ", "Deepfakes are AI-generated videos or audio that mimic real people. They’re designed to look and sound convincing, often used to spread misinformation or commit fraud by faking someone’s voice or face. That’s what makes them dangerous, and effective. "],
Core Practices for Spotting Deepfakes
[98, "Core Practices for Spotting Deepfakes ", "Core Practices for Spotting Deepfakes "],
Protecting yourself against malicious deepfakes requires you to understand how deepfakes are created, how they’re used, and how you can spot the subtle signs of video or audio manipulation. Over the coming pages, we’ll explore these concepts further.
[99, "Protecting yourself against malicious deepfakes requires you to understand how deepfakes are created, how they’re used, and how you can spot the subtle signs of video or audio manipulation. Over the coming pages, we’ll explore these concepts further.", "Protecting yourself against malicious deepfakes requires you to understand how deepfakes are created, how they’re used, and how you can spot the subtle signs of video or audio manipulation. Over the coming pages, we’ll explore these concepts further."],
Recognizing Manipulated Videos
[100, "Recognizing Manipulated Videos ", "Recognizing Manipulated Videos "],
Recognizing manipulated videos is a critical part of deepfake awareness. Even the most convincing videos can contain subtle flaws that reveal they’ve been altered. This includes unusual facial movements, mismatched lighting, unnatural speech patterns, or behaviors that seem out of character for the person.
[101, "Recognizing manipulated videos is a critical part of deepfake awareness. Even the most convincing videos can contain subtle flaws that reveal they’ve been altered. This includes unusual facial movements, mismatched lighting, unnatural speech patterns, or behaviors that seem out of character for the person. ", "Recognizing manipulated videos is a critical part of deepfake awareness. Even the most convincing videos can contain subtle flaws that reveal they’ve been altered. This includes unusual facial movements, mismatched lighting, unnatural speech patterns, or behaviors that seem out of character for the person. "],
Which of the following is a possible sign that a video may be a deepfake?
[102, "Which of the following is a possible sign that a video may be a deepfake? ", "Which of the following is a possible sign that a video may be a deepfake? "],
Out-of-sync voice and mouth movements
[103, "Out-of-sync voice and mouth movements ", "Out-of-sync voice and mouth movements "],
Perfectly natural facial expressions
[104, "Perfectly natural facial expressions ", "Perfectly natural facial expressions "],
Lighting and shadows that match the environment perfectly
[105, "Lighting and shadows that match the environment perfectly ", "Lighting and shadows that match the environment perfectly "],
Smooth, natural speech with emotional tone
[106, "Smooth, natural speech with emotional tone", "Smooth, natural speech with emotional tone"],
Deepfake videos often struggle to perfectly match lip movements with speech, creating a slight delay or mismatch.
[107, "Deepfake videos often struggle to perfectly match lip movements with speech, creating a slight delay or mismatch.", "Deepfake videos often struggle to perfectly match lip movements with speech, creating a slight delay or mismatch."],
Question Authenticity
[108, "Question Authenticity", "Question Authenticity"],
Never take visual or audio content at face value, especially when the stakes are high. Treat unexpected, urgent, or sensitive requests with caution and investigate further before responding.
[109, "Never take visual or audio content at face value, especially when the stakes are high. Treat unexpected, urgent, or sensitive requests with caution and investigate further before responding. ", "Never take visual or audio content at face value, especially when the stakes are high. Treat unexpected, urgent, or sensitive requests with caution and investigate further before responding. "],
Verify Identities
[110, "Verify Identities ", "Verify Identities "],
Always confirm the person’s identity using known secure communication channels. Call them directly, arrange a face-to-face meeting, or use a previously verified contact method before acting on any request.
[111, "Always confirm the person’s identity using known secure communication channels. Call them directly, arrange a face-to-face meeting, or use a previously verified contact method before acting on any request. ", "Always confirm the person’s identity using known secure communication channels. Call them directly, arrange a face-to-face meeting, or use a previously verified contact method before acting on any request. "],
Which is the safest response to an unexpected request for sensitive information on a video call?
[112, "Which is the safest response to an unexpected request for sensitive information on a video call?", "Which is the safest response to an unexpected request for sensitive information on a video call?"],
Share the information immediately
[113, "Share the information immediately", "Share the information immediately"],
Verify the request using a trusted communication channel
[114, "Verify the request using a trusted communication channel", "Verify the request using a trusted communication channel"],
Record the call for later review
[115, "Record the call for later review ", "Record the call for later review "],
Ask them security questions you just made up on the spot
[116, "Ask them security questions you just made up on the spot ", "Ask them security questions you just made up on the spot "],
Deepfakes and impersonation scams can make video calls appear genuine. Visual and audio cues aren’t always enough to confirm someone's identity. If the stakes are high, verify someone’s identity through a secure, trusted communication channel.
[117, "Deepfakes and impersonation scams can make video calls appear genuine. Visual and audio cues aren’t always enough to confirm someone's identity. If the stakes are high, verify someone’s identity through a secure, trusted communication channel. ", "Deepfakes and impersonation scams can make video calls appear genuine. Visual and audio cues aren’t always enough to confirm someone's identity. If the stakes are high, verify someone’s identity through a secure, trusted communication channel. "],
Which of the following is NOT a key component of deepfake awareness?
[118, "Which of the following is NOT a key component of deepfake awareness?", "Which of the following is NOT a key component of deepfake awareness?"],
Recognizing manipulation
[119, "Recognizing manipulation ", "Recognizing manipulation "],
Questioning authenticity
[120, "Questioning authenticity", "Questioning authenticity"],
Verifying identities
[121, "Verifying identities", "Verifying identities"],
Learning to create realistic deepfakes
[122, "Learning to create realistic deepfakes", "Learning to create realistic deepfakes"],
Deepfake awareness is about protecting yourself from manipulation, not creating it. The focus is on spotting fakes, questioning what you see or hear, verifying authenticity, and staying informed.
[123, "Deepfake awareness is about protecting yourself from manipulation, not creating it. The focus is on spotting fakes, questioning what you see or hear, verifying authenticity, and staying informed.", "Deepfake awareness is about protecting yourself from manipulation, not creating it. The focus is on spotting fakes, questioning what you see or hear, verifying authenticity, and staying informed."],
Seeing or hearing someone in a video call is always enough to confirm their identity.
[124, "Seeing or hearing someone in a video call is always enough to confirm their identity.", "Seeing or hearing someone in a video call is always enough to confirm their identity."],
Deepfakes can mimic a person’s face and voice in real time, making sight and sound alone unreliable for identity verification.
[125, "Deepfakes can mimic a person’s face and voice in real time, making sight and sound alone unreliable for identity verification. ", "Deepfakes can mimic a person’s face and voice in real time, making sight and sound alone unreliable for identity verification. "],
Deepfake technology is only a risk to celebrities and public figures.
[126, "Deepfake technology is only a risk to celebrities and public figures.", "Deepfake technology is only a risk to celebrities and public figures."],
Scammers increasingly target everyday individuals, colleagues, and family members using deepfakes to exploit trust.
[127, "Scammers increasingly target everyday individuals, colleagues, and family members using deepfakes to exploit trust. ", "Scammers increasingly target everyday individuals, colleagues, and family members using deepfakes to exploit trust. "],
Why Deepfake Awareness Matters
[128, "Why Deepfake Awareness Matters ", "Why Deepfake Awareness Matters "],
Deepfake scams can lead to financial loss, reputational damage, data breaches, and erosion of trust. By staying alert and applying deepfake awareness practices, individuals and organizations can prevent costly mistakes and protect their integrity.
[129, "Deepfake scams can lead to financial loss, reputational damage, data breaches, and erosion of trust. By staying alert and applying deepfake awareness practices, individuals and organizations can prevent costly mistakes and protect their integrity.", "Deepfake scams can lead to financial loss, reputational damage, data breaches, and erosion of trust. By staying alert and applying deepfake awareness practices, individuals and organizations can prevent costly mistakes and protect their integrity."],
Which of the following is a potential consequence of failing to spot a deepfake scam?
[130, "Which of the following is a potential consequence of failing to spot a deepfake scam? ", "Which of the following is a potential consequence of failing to spot a deepfake scam? "],
Increased trust in digital communications
[131, "Increased trust in digital communications ", "Increased trust in digital communications "],
Financial loss, reputational harm, and data compromise
[132, "Financial loss, reputational harm, and data compromise ", "Financial loss, reputational harm, and data compromise "],
More efficient workflow and reduced security checks
[133, "More efficient workflow and reduced security checks", "More efficient workflow and reduced security checks"],
Improved AI training data
[134, "Improved AI training data ", "Improved AI training data "],
Falling for a deepfake scam can have serious consequences, including money loss, stolen information, and damaged trust.
[135, "Falling for a deepfake scam can have serious consequences, including money loss, stolen information, and damaged trust.", "Falling for a deepfake scam can have serious consequences, including money loss, stolen information, and damaged trust."],
Eye Movement
[136, "Eye Movement", "Eye Movement"],
Eyes are hard to fake. In real footage, blinks are irregular, gaze shifts, and pupils react to light. Deepfakes miss these cues and may have missing or mechanical blinks, fixed stares, over-bright eyes or static pupils. Take note of the deepfake video below, where the speaker doesn't blink.
[137, "Eyes are hard to fake. In real footage, blinks are irregular, gaze shifts, and pupils react to light. Deepfakes miss these cues and may have missing or mechanical blinks, fixed stares, over-bright eyes or static pupils. Take note of the deepfake video below, where the speaker doesn't blink.", "Eyes are hard to fake. In real footage, blinks are irregular, gaze shifts, and pupils react to light. Deepfakes miss these cues and may have missing or mechanical blinks, fixed stares, over-bright eyes or static pupils. Take note of the deepfake video below, where the speaker doesn't blink."],
Which eye behavior is a strong red flag for a deepfake?
[138, "Which eye behavior is a strong red flag for a deepfake?", "Which eye behavior is a strong red flag for a deepfake?"],
Irregular, natural blinking
[139, "Irregular, natural blinking", "Irregular, natural blinking"],
No blinking for extended periods
[140, "No blinking for extended periods ", "No blinking for extended periods "],
Brief glances off-camera
[141, "Brief glances off-camera", "Brief glances off-camera"],
Pupils constrict when light increases
[142, "Pupils constrict when light increases", "Pupils constrict when light increases"],
Humans blink irregularly, especially during conversation. Deepfakes often suppress or mistime blinks because generators prioritize face/voice alignment and struggle with involuntary micro-movements. The result is long, fixed stares or robotic blink bursts.
[143, "Humans blink irregularly, especially during conversation. Deepfakes often suppress or mistime blinks because generators prioritize face/voice alignment and struggle with involuntary micro-movements. The result is long, fixed stares or robotic blink bursts.", "Humans blink irregularly, especially during conversation. Deepfakes often suppress or mistime blinks because generators prioritize face/voice alignment and struggle with involuntary micro-movements. The result is long, fixed stares or robotic blink bursts."],
Lip Sync Mismatch
[144, "Lip Sync Mismatch", "Lip Sync Mismatch"],
Deepfakes tend to struggle when it comes to lip-syncing. Real speech syncs lip, jaw, and cheek movement with audio. Deepfakes often have a slight lag, wrong mouth shapes, or mouth movement that doesn’t line up.
[145, "Deepfakes tend to struggle when it comes to lip-syncing. Real speech syncs lip, jaw, and cheek movement with audio. Deepfakes often have a slight lag, wrong mouth shapes, or mouth movement that doesn’t line up.", "Deepfakes tend to struggle when it comes to lip-syncing. Real speech syncs lip, jaw, and cheek movement with audio. Deepfakes often have a slight lag, wrong mouth shapes, or mouth movement that doesn’t line up."],
Which sign suggests a video may be a deepfake?
[146, "Which sign suggests a video may be a deepfake? ", "Which sign suggests a video may be a deepfake? "],
Natural head movement and facial expressions
[147, "Natural head movement and facial expressions", "Natural head movement and facial expressions"],
Out-of-sync voice and mouth movements
[148, "Out-of-sync voice and mouth movements ", "Out-of-sync voice and mouth movements "],
Smooth camera focus
[149, "Smooth camera focus", "Smooth camera focus"],
Consistent lighting across the scene
[150, "Consistent lighting across the scene", "Consistent lighting across the scene"],
Real speech has tight timing between sounds and mouth shapes. Deepfakes often miss that micro-timing, so lips lag the audio or form the wrong shapes.
[151, "Real speech has tight timing between sounds and mouth shapes. Deepfakes often miss that micro-timing, so lips lag the audio or form the wrong shapes. ", "Real speech has tight timing between sounds and mouth shapes. Deepfakes often miss that micro-timing, so lips lag the audio or form the wrong shapes. "],
Background & Environmental Mismatches
[152, "Background & Environmental Mismatches ", "Background & Environmental Mismatches "],
Audio–visual mismatch is a deepfake tell. If the voice sounds studio-clean while the person is outdoors, if echo appears without a source, or clothes ripple with silent wind, something’s off.
[153, "Audio–visual mismatch is a deepfake tell. If the voice sounds studio-clean while the person is outdoors, if echo appears without a source, or clothes ripple with silent wind, something’s off.", "Audio–visual mismatch is a deepfake tell. If the voice sounds studio-clean while the person is outdoors, if echo appears without a source, or clothes ripple with silent wind, something’s off."],
What’s the most telling sign that this video is an AI-generated deepfake?
[154, "What’s the most telling sign that this video is an AI-generated deepfake?", "What’s the most telling sign that this video is an AI-generated deepfake?"],
The lighting is consistent across the scene
[155, "The lighting is consistent across the scene ", "The lighting is consistent across the scene "],
The subject’s facial expressions seem rigid or unnatural
[156, "The subject’s facial expressions seem rigid or unnatural", "The subject’s facial expressions seem rigid or unnatural"],
A passing bus briefly blocks the subject
[157, "A passing bus briefly blocks the subject", "A passing bus briefly blocks the subject"],
The mouth movements are perfectly in sync with the voice
[158, "The mouth movements are perfectly in sync with the voice", "The mouth movements are perfectly in sync with the voice"],
Deepfakes often struggle to replicate subtle human expressions. While a face might blink or smile, it can still look off, with expressions that are stiff, delayed, or robotic. This lack of nuance is a strong giveaway. Real human emotion involves micro-expressions and muscle movements that are hard for AI to fake convincingly.
[159, "Deepfakes often struggle to replicate subtle human expressions. While a face might blink or smile, it can still look off, with expressions that are stiff, delayed, or robotic. This lack of nuance is a strong giveaway. Real human emotion involves micro-expressions and muscle movements that are hard for AI to fake convincingly. ", "Deepfakes often struggle to replicate subtle human expressions. While a face might blink or smile, it can still look off, with expressions that are stiff, delayed, or robotic. This lack of nuance is a strong giveaway. Real human emotion involves micro-expressions and muscle movements that are hard for AI to fake convincingly. "],
The Importance of Deepfake Awareness
[160, "The Importance of Deepfake Awareness ", "The Importance of Deepfake Awareness "],
By recognising manipulation, questioning authenticity, verifying identities, and staying informed, you can reduce the risk of falling victim to deepfake scams. Awareness today means protection tomorrow, for you and your organization.
[161, "By recognising manipulation, questioning authenticity, verifying identities, and staying informed, you can reduce the risk of falling victim to deepfake scams. Awareness today means protection tomorrow, for you and your organization.", "By recognising manipulation, questioning authenticity, verifying identities, and staying informed, you can reduce the risk of falling victim to deepfake scams. Awareness today means protection tomorrow, for you and your organization."],
[162, "", ""],
Original Text (English)
Correct Translation (English)
Submissions are reviewed and applied within 48 hours.
Color Theme And Background Selector
×
Colorful Theme
Select Theme
Blue Theme
Select Theme
Light Theme
Select Theme
Purple Theme
Select Theme
Green Theme
Select Theme
Dark Theme
Select Theme
Note: The Colorful, Blue, Light, and Dark Themes are all WCAG 2.2 Level AA conformant.
