The Dangers of AI on Social Media Nobody Talks About

Five years ago, the worst thing on your feed was a bad filter. Today it might be a person who doesn't exist, saying something that never happened, in a place they've never been, with ten thousand bot comments agreeing with them.

That's not a sci-fi future. It's your feed right now. Here are the dangers of AI on social media that matter most, in plain English, and what you can actually do about them.

1. You can't trust your own eyes anymore

AI video tools can now create realistic faces, voices and scenes in minutes. Researchers testing the newest video models in 2026 found that both people and detection tools were close to a coin flip at telling real clips from fake ones. When experts can't tell, a quick scroll on your phone doesn't stand a chance.

The result: a video of a fight, a crash, a celebrity or a friend can be completely made up, and it will look exactly as real as the truth.

2. Bots are a huge share of the internet

Cloudflare reported in 2026 that AI traffic passed human traffic on the web for the first time. On social apps, that shows up as bot accounts, bot comments, bot likes and bot followers. When a post looks popular, you can't assume real people made it popular.

3. "AI slop" is drowning out real people

AI slop is the flood of low-effort, auto-generated posts built only to farm views. It's cheap to make, so there's endless amounts of it, and it competes with real creators for the same attention. LinkedIn even added a "seems like AI slop" button in 2026, and over a million people clicked it.

4. Fake influencers and fake reviews

Entire influencer accounts are now AI-generated: the face, the lifestyle, the "candid" photos. Brands pay them. Followers trust them. Nothing about them is real, and most people never find out.

5. Scams that look and sound like people you know

Voice cloning and face-swapping make it easy to impersonate a friend, a family member or a boss. The FTC and security companies keep warning about "urgent" videos or voice messages that ask for money. The better the fake, the easier the scam.

6. The "dead internet" feeling

Put it all together and you get what people call the dead internet theory: the sense that most of what you see online isn't made by humans anymore. Whether or not the theory is fully true, the feeling is real, and it's why so many people are tired of scrolling.

What you can do about it

• Slow down on shocking videos. The more emotional a clip makes you, the more you should check it before sharing.
• Check the source. Who posted it first? How old is the account? What else have they posted?
• Use apps that make faking harder. The strongest fix isn't a better detector. It's a platform where fake content can't get in to begin with.

How Portal handles AI

Portal is a social app built on one simple rule: every post has to be real.

• Camera-only posting. Photos and videos are captured live in the app. You can't upload from your camera roll, so AI-generated clips can't be posted.
• Location-based. Posts are tied to where they were actually taken, so you see what's happening in your city from people who are actually there.
• No AI content, no bots. Portal is for humans only.
• 48-hour posts. Content is in-the-moment, then saves to your profile as memories.

AI isn't going away. But you can choose where you spend your time, and who you trust.

Download Portal free on iPhone

Frequently asked questions