AI Misinformation and Deepfakes

Learning Objectives

• Recognize common signs of AI-generated text, images, and video
• Use practical verification techniques to check suspicious content
• Build information hygiene habits for the age of AI

The New Reality

AI has made it dramatically easier to create convincing fake content. A few years ago, creating a realistic fake video required Hollywood-level resources. Today, anyone with an internet connection can generate:

• Fake images that look like real photographs
• Fake videos of real people saying things they never said (deepfakes)
• Fake articles written in convincing journalistic style
• Fake audio that sounds like a specific person's voice
• Fake reviews, comments, and social media posts at massive scale

This is not a theoretical problem. In 2025 and 2026, AI-generated content has been used in political campaigns, financial scams, corporate fraud, and personal harassment. You have almost certainly encountered AI-generated content without realizing it.

Spotting AI-Generated Text

AI-written text has become very hard to detect, but there are patterns to watch for:

Common tells:

• Overly balanced and hedging. AI tends to present both sides of everything, even when one side is clearly stronger. Phrases like "on the other hand" and "it is important to note" appear frequently.
• Perfect grammar with no personality. Human writing has quirks, incomplete sentences, and personal voice. AI writing is technically correct but often feels generic.
• Confident but unsourced claims. AI states facts without citations. If an article makes specific claims with no links or references, be skeptical.
• Repetitive structure. AI often follows a predictable pattern: introduce topic, list points, conclude with summary. Real writing is messier.

What does NOT work:

• AI detection tools (like GPTZero) have high false positive rates — they often flag human writing as AI-generated. Do not rely on them as definitive proof.

Spotting AI-Generated Images

AI image generation has improved rapidly, but imperfections remain:

Check for:

• Hands and fingers. AI still struggles with hands — look for extra fingers, merged fingers, or unnatural poses. This is improving but still the most reliable tell in 2026.
• Text in images. AI-generated text within images often has spelling errors, nonsensical words, or inconsistent fonts.
• Background inconsistencies. Look at the edges of the main subject. Objects may blend unnaturally, straight lines may waver, or patterns may not repeat correctly.
• Skin and hair. AI can produce an "uncanny valley" smoothness to skin, or hair that merges with the background in unnatural ways.
• Symmetry. AI often produces faces that are more symmetrical than real human faces. Earrings, glasses, and collars may not match on both sides.

Tools that help:

• Reverse image search (Google Images, TinEye) — if the image exists nowhere else online, it may be AI-generated
• Metadata inspection — real photos contain EXIF data (camera model, date, location). AI images typically have no metadata or generic metadata

Spotting Deepfake Videos

Video deepfakes are harder to spot but have tells:

• Lip sync. The mouth movements may not perfectly match the audio, especially on consonant sounds
• Blinking. Early deepfakes had unnatural blinking patterns. Newer ones are better, but watch for moments where the face seems to "reset"
• Edge artifacts. Look at the boundary between the face and hair/background. You may see flickering, blurring, or color shifts
• Lighting consistency. The lighting on the face may not match the lighting in the rest of the scene
• Context. Ask: does it make sense that this person would say this, in this setting, to this audience? If a video seems designed to provoke an extreme reaction, be extra skeptical

The Verification Toolkit

When you encounter suspicious content, use this three-step process:

1. Check the Source

Where did this content come from? A major news outlet with editorial standards? A random social media account? An anonymous website? The source tells you more than the content itself.

2. Cross-Reference

Search for the same claim on other reputable sources. If a dramatic claim is only reported by one source, that is a red flag. Use specific search queries: paste a key quote into Google in quotation marks.

3. Check the Timeline

Misinformation often involves real events with false context — a real photo from a different time or place, a real quote taken out of context. Check when and where the original content was created.

Building Information Hygiene

Like hand washing for your information diet:

1. Diversify your sources. Do not rely on a single platform or outlet for news
2. Be skeptical of perfection. If a photo, video, or article seems too perfect, too outrageous, or too convenient for a particular narrative — investigate
3. Check dates. Old content recirculated as new is a common misinformation tactic
4. Follow the money. Who benefits from you believing this? That question cuts through a lot of noise
5. Teach your kids. If you have children, these verification habits are essential life skills

Key Takeaways

• AI-generated fake content is getting harder to detect — no single technique catches everything
• For text: watch for over-hedging, generic voice, and unsourced claims
• For images: check hands, text, backgrounds, and use reverse image search
• For video: watch lip sync, blinking, face edges, and ask if the context makes sense
• The best defense is the verification mindset: check the source, cross-reference, check the timeline
• Slow down before sharing — emotional reactions are how misinformation spreads