AI impersonation is here. Until now, we could always say “I saw it. I heard it. I know it.” From today on, we simply cannot believe everything that we see and hear. And if truth is still important to us (which I hope it is), then we are going to need to know how to parse through the universe of muck.
So, what is a deepfake?
A deepfake is a piece of content — audio, video or graphic — that is digitally generated (almost always with artificial intelligence these days) and it represents something that never actually happened but is frequently used as a weapon to alter behavior or belief — e.g. the voter suppression case in February where Texas based Life Corporation sent a deepfake robocall of President Biden’s voice telling voters that they don’t have to vote in the New Hampshire primary election because, “it’s important you save your votes for the November election.”
Or from January when a teacher in Baltimore retaliated against his principal by creating a voice recording deepfake of him making racial and antisemitic comments about community members and students. Once leaked over social media, this caused immediate community outrage to the point where he was put on unpaid leave, received death threats and required 24/7 police protection for over a month until Baltimore County authorities announced the recording was a fake and arrested Dazhon Darien a few weeks ago charges of stalking, theft, disruption of school operations and retaliation against a witness.
Perhaps libel and slander (defamation) could play a role here, but not always. In almost half of the states, defamation can be charged as a crime (and this has been repealed here and there). In California, defamation is a civil matter and limited to their right to recover damages (think Johnny Depp v. Amber Heard from 2022). In Maryland, common-law defamation was repealed in 2013. While generally a good thing due to its historically broad overreach, this doesn’t help Principal Eiswert.
Why do deepfakes even exist?
There are a few reasons, and I’ll be short with the explanations but know that they each merit much deeper conversation:
• Humans overestimate how good they are at knowing when they are looking at a deepfake. Is this you? Try Northwestern University’s test.
• It’s got the best marketing built in. The lines and imagery were crafted to perfection, intended to trigger you in just the right way for just the right reason. Even tap into your most well-meaning intentions, like making sure your vote counts so you skip voting. I don’t know, you’re probably pretty busy like everyone else so it’s certainly possible that you missed a change in voting procedure, so how great of President Biden to stump for your candidate and get out the “right” vote, right?
Recommended for you
• Seeing it is believing it. Cognitively, seeing it with our own eyes makes it much harder to think that what we are seeing or hearing is not real. How could it not be real? You saw it yourself, right? In the real world, this still remains mostly true. But online and in spaces where digital content can play any role, it no longer is.
How do you know it’s a deepfake?
Well, technology is getting better every day, so it’s hard. But here are a few tips:
• Look for inconsistencies in the image: blurriness where there shouldn’t be any, skin tone or shadowing that looks off, unnatural blinking or twitches in the cheeks and lips when speaking or smiling, background imagery that doesn’t match the foreground, or inconsistencies in the shading in various parts of the image.
• Look for inconsistencies in the voice recording: Are any words cut off anywhere? Are the inflections just a little bit weird for what that person is saying (This happens when people splice together words from different audio recordings)?
• Ask yourself: Would they say that like that? Because the wording is so “perfectly crafted,” it’s also probably going to be too perfect. What would that person actually say if they were intending to convey that piece of information?
Long and short — double, triple, quadruple fact check anything that you see online that makes you think twice. That gives you pause. That is a little bit surprising or shocking. That seems counter intuitive. Even something that seems to reinforce a gut feeling you had but also thought you were a little bit crazy for thinking it in the first place. Fact check everything before you commit it to belief and pass it on. We simply all need to do this to get through this.
Annie Tsai is chief operating officer at Interact (tryinteract.com), early stage investor and advisor with The House Fund (thehouse.fund), and a member of the San Mateo County Housing and Community Development Committee. Find Annie on Twitter @meannie.

(1) comment
The better AI gets at creating deepfakes, the harder it will be to detect them. Chicolini while deep-faking Rufus T. Firefly asked, "Well, who you gonna believe, me or your own eyes?"
Welcome to the discussion.
Log In
Keep the discussion civilized. Absolutely NO personal attacks or insults directed toward writers, nor others who make comments.
Keep it clean. Please avoid obscene, vulgar, lewd, racist or sexually-oriented language.
Don't threaten. Threats of harming another person will not be tolerated.
Be truthful. Don't knowingly lie about anyone or anything.
Be proactive. Use the 'Report' link on each comment to let us know of abusive posts.
PLEASE TURN OFF YOUR CAPS LOCK.
Anyone violating these rules will be issued a warning. After the warning, comment privileges can be revoked.