Ever since generative AI took to image and video generation, the internet is brimming with what's known as "AI slop". You may come across it while scrolling through YouTube shorts, Instagram Reels or through X threads. Before more sophisticated image models arrived, these used to be low-effort, easily identifiable content, but now, generative AI has advanced to a point that some media blurs the lines between fact and fiction.

Navigating the internet in the age of gen AI certainly requires a robust toolkit for verification to protect yourself from disinformation, scams, and fraud. While Grok handles real-time information checks within the ecosystem of X, the challenge remains when moving across the broader web. Outside this bubble, I rely on Gemini. Here's why it has become indispensable in my fact-checking arsenal, and why Gemini is probably the best when it comes to the job.

Gemini is the definitive 'slop' safeguard

It benefits from its real-time web integration

Generative models have come a long way since ChatGPT 3.0, but they can easily slip into the "hallucination" territory when you least expect it. Google's LLM has a perfect safeguard against this, and it benefits directly from its integration with the search engine and the real-time web data it draws inferences from.

Gemini's Double-check tool feels reliable against model hallucinations, and it doesn't matter whose output you are checking. The responses entered are subject to a variety of cross-references indexed against the web, and you're provided direct sources that either confirm or negate the data points. I also find it particularly useful when I am suspecting potential 'sycophancy', a phenomenon when every assertion I make is met with overwhelming affirmation.

AI-generated images have become very convincing

And Gemini's SynthID is the perfect remedy for it

You've probably seen something bizarre online that's almost as absurd as your friend posting a selfie with Barack Obama at their local Walmart. In the past, you'd look for obvious signs of image generation, like a sixth finger or two left hands, but generative models like Nano Banana produce images that are so organic they bypass the "uncanny valley" almost entirely. You can't find a jagged pixel or oddly melting shadows to tip you off, because everything looks as real as a polaroid.

The problem here is that not every generated image is used for a harmless prank. AI-induced disinformation carries the capability to shake belief systems and manipulate public discourse, and that's why SynthID is a critical guardrail for me. The SynthID technology was developed by Google DeepMind as a back-end verification tool which is now available for all users. The reason why it's so good at catching AI-generated images is because it not only scans the media's metadata, but also reads for the otherwise imperceptible digital watermarks which are directly integrated into pixels or audio waves of content generated by the Nano Banana model. Additionally, Gemini is also good at catching images generated by other models such as Stable Diffusion, ChatGPT 5.0 and Grok through metadata analysis.

How to use Gemini's SynthID feature

And never be fooled by AI slop again

Using the synthID feature is as straightforward as typing in a prompt. To use Gemini's verification system, simply upload any suspicious media, such as an image, an audio or a video file directly into the chat. Next, you can either use a prompt such as, "Was this image generated by AI?" or simply type in, "#SynthID". This activates the scanning protocol, prompting Gemini to check for the watermarks and confirm whether all or part of the image was generated with AI.

What I've noticed is, even if a watermark isn't found, the tool goes further and provides a breakdown of any aberrations or visual artefacts that give away signs of generative AI. For example, in one sample that I ran, the model was able to check the image's metadata and confirm that the image in question was likely generated with Stable Diffusion.

This feature, of course, works well only if you pair it with your intuition. Despite their sophistication, most generated media will come with signs of 'synthesis', even if they're not as obvious as they used to be a couple of years ago. If something depicted or said seems wildly out of place, it most likely warrants a check, and that's where Gemini helps.

It's important to protect your feed from disinformation

Synthetic content is becoming indistinguishable from reality, and staying skeptical online is almost necessary. Fact-checking tools provide you with a great way to protect yourself from disinformation, but it's worth remembering that they're not always infallible, and as models continue to evolve, the gaps between what's organic and what's designed to deceive will only become more narrow.