Would you believe if I told you that more than 15 billion AI-generated images are already circulating online? OpenAI’s latest data shows that users create more than 2 million images daily with DALL-E 2, an AI art creator.

These data points to one thing- technological advancements have opened the floodgates for all kinds of fake images generated with artificial intelligence (AI). Daily, our social media feeds are flooded with a new type of spam, often referred to as “AI slop,” designed to generate clicks and interactions.

Interestingly, most of this content is absurd and unrealistic, with users creating impossible images. However, as AI technology continues to develop at a breakneck speed, these images are improving in quality, becoming increasingly realistic.

But How Can You Tell if an Image Is Artificially Generated?

The devil is in the details, especially when it comes to humans. AI often struggles with generating hands, frequently producing fingers that are too long, too short, deformed, or incorrectly numbered, sometimes resulting in images with four, six, or even seven fingers.
Additionally, eyes often appear misaligned, feature inconsistent reflections, or look lifeless despite the overall realism of the image.

Watch Out for Distorted Text

One of the clearest giveaways of an AI-generated image is its inability to produce readable text. AI models frequently struggle to generate coherent words. This issue is particularly noticeable on signs, billboards, and product labels, where the text often appears as gibberish. For instance, AI-generated posters could feature words like “encherining”, and “cartchy tuns” instead of presumably “enchanting” and “catchy tunes.”

Misaligned Shadows and Unnatural Reflections

Lighting and shadows are key indicators of an image’s authenticity. In real-world photography, light follows predictable patterns, casting natural shadows and reflections. However, AI-generated images often fail to replicate this accurately, producing reflections and shadows that don’t align with the light source, giving them an eerie, uncanny quality.

Another common issue is excessively uniform lighting. AI-generated visuals often seem too smooth, glossy, or polished, missing the natural depth and variation found in real photographs. This can lead to an unnatural, almost “plastic” or “Disney-like” aesthetic. In addition to lighting inconsistencies, AI-generated images often lack the subtle imperfections that provide real-world photography with its authenticity.

Unnaturally Flawless Features

AI tends to over-smooth textures, making skin, surfaces, and fabric appear unnaturally flawless. Another telltale sign of artificially created content is repetitive patterns. AI struggles to generate truly distinct elements, often leading to cloned objects or subtly altered repeating details.

It is important to pay close attention to elements like crowds, which may feature individuals with nearly identical faces, or background patterns with blatant repetitions.

AI-generated content is no longer a futuristic concern, it has successfully crept into Ghana’s information space. During the December elections, AI was actively used to create misleading political content, shaping narratives and influencing public opinion. This highlights the urgent need for enhanced media and information literacy (MIL) education to equip citizens with the critical skills to identify and combat misinformation. The onus lies on all of as stakeholders and individuals to remain vigilant and ensure that technology serves the truth, not deception.

Below is a quick checklist for spotting AI-generated images:

• Check fingers and eyes for abnormalities
• Look for distorted or unreadable text
• Observe lighting inconsistencies and unnatural reflections
• Look for overly flawless and repetitive textures or patterns
• Notice unrealistic perfection or unnatural smoothness

Tools and Platforms for Image Verification

• Google Reverse Image Search:- Quickly check if an image has appeared elsewhere online.
• TinEye:- A reverse image search tool specialized in tracking the origin of images.
• FotoForensics:- Offers digital forensic analysis of images, highlighting potential edits or manipulations.
• InVID Verification Plugin:- Browser plugin that verifies images and videos, particularly useful for social media content.