7 Tips on How to Spot AI Images

The background. Are people in the background looking at the unusual thing going on? If they are going about their business, it is likely a fake. Often the background of AI images will be distorted. 

Other video and photos. If you don’t see other video from different angles, you may be looking at AI generated images. It would be unlikely that there would be only a single image or video of something odd or newsworthy. Don’t share until there is confirmation.

Details. AI generators are not good at details—like fingers, hands and hair. Often AI software will show too many fingers or odd hand placement.* 

Writing. Look closely at writing on a sticker, street sign or billboard. Often it is blurry when it shouldn’t be or else the letters are formed wrong, or the letters don’t form words.

Glossy. The overly-glossy look, similar to some stock photos, can be an AI giveaway. Watch for people with plastic-looking faces.

Source. Is the person or organization sharing the image reliable?

The Eyes. By using methods conventionally used to "measure the shapes of galaxies," researchers have found that deepfake images don't have the same consistency in reflections across both eyes. However, this method of detection does have false positives and false negatives.

*Rice University research could make weird AI images a thing of the past

23 AI-based text/image/video creation tools

ArtBreeder
Create portraits and landscape images with this AI-based creation tool. Free. An introduction and review here.

ChatGPT
This OpenAI chatbot remembers what you've written or said, so the interaction has a dynamic conversational feel. Give the software a prompt and it creates articles, even poetry. It writes code, too. And explains the code. Doesn’t do legit sources and limited to info from before 2022. Free.

Co-Pilot
Released by Microsoft-owned GitHub built on OpenAI technology that can translate basic human instructions into functional computer code. Intended for developers.

Copy.ai
AI text-generation tool.

Craft.do
Document-creation tool with AI features.

DALL-E
OpenAI’s tool that turns written text into realistic images using AI. Named after painter Salvador Dali and Disney Pixar’s WALL-E. A limited number of images are free.

Google Pinpoint
This tool uses AI to analyze PDFs, strip text from images and transcribe audio.

GPT-2 Detector
OpenAI’s GPT-2 Detector (Hugging Face) is a tool that helps to identify AI generated text.

GPTZero
Detects whether an essay was written by ChatGPT to help educators to combat AI-based plagiarism Built by a computer science student at Princeton who was a former data journalist with the BBC.

Jasper AI
AI story writing tool for fiction and nonfiction. Pick a tone of voice for style. Pre-built templates available. However, no sources are provided. $29 month.

Lensa
Create digital self-portraits, made with AI technology through the open source Stable Diffusion model that renders selfies into artwork. Developed by Prisma Labs, locataed in California by Russian developers. One week free trial then $30 a month. The avatar tool costs a separate $3.99 for 50 images.

Lex
AI text-generation tool.

Make-A-Video
Meta’s AI system that turns text into video. Not yet available to the public.

Maker
Generates written and visual content. Free trial then $25 a month.

MidJourney
This AI image generator uses machine learning to create pictures based on text. Created a picture that sparked a controversy by winning an art competition.

Munch
Uses AI to repurpose video content for social channels and more.

NightCafe
Create art with the help of Artificial Intelligence.

NovelAI
AI story creator. Easy-to-use but fantasy fiction only. Starts at $10 a month.

Postwise.ai
AI-driven Twitter writing tool

QuillBot
An AI-driven writing tool that paraphrases what you say.

Stable Diffusion
Generates visual creations through AI. Since it is open-sourced, anyone can view the code. Fewer restrictions on how it can be used than DALL-E.

VanceAI Art Generator
Read more about it here.

VALL-E
Microsoft's AI-powered can replicate someone’s voice with just 3-second sample. The voice synthesis machine learning model is not yet available to the public.

More Tech Tools