Oriol Salvador

AI-generated YouTube thumbnails are ugly

AI can generate an optimized YouTube thumbnail in seconds. But if every channel uses the same visual formula, how can anyone stand out?

Out of all the blog posts sharing my expertise on YouTube channel management and optimisation that I have published here since 2019, there’s one that still drives a lot of traffic to this day. It’s the one where I share best practices to create eye-catching YouTube thumbnails.

I love all my blog posts equally, but I’m particularly proud of that one. The recommendations I share still hold true and there are elements on it, like a graphic to show the recommended size of a YouTube thumbnail and the green and red zones to structure visual elements on it, that still apply today. According to my web analytics, that blog post has reached far and wide online.

Fun fact: Google Search Console shows me this blog post appears in many search results just because of the use of the word porn.

If you are new here, I’m a journalist and digital media editor with a lot of experience managing and optimising video content for YouTube.

I’ve worked with musicians, filmmakers, journalists, dancers, music labels, festivals, distributors, news outlets, non-profit organizations… Name it, I’ve probably worked on a video about it. I did a webinar once, too.

In this work, there’s a browser extension I use a lot, it’s called vidIQ. I use it because it surfaces the metadata of your videos, and videos from others, so you can find the best keywords for your video, for example. Nowadays, it relies heavily on artificial intelligence to offer a series of AI-powered features. Among them, there’s the one I want to talk about today.

Human curation against AI

AI is generating more music than ever, making human curation essential to find what is actually worth your attention.

It’s Friday. You wake up and open your music streaming platform of choice to listen to new releases this week. They come in an automated playlist generated based on your listening habits and favourite music.

You put it on shuffle and, after a couple of songs, one plays where you don’t recognize the voice, but sounds generic and, let’s be frank, not great.

When you check the details you recognize the band or artist’s name, but the album cover screams ‘generated by ChatGPT’ and the song was, in fact, also generated by artificial intelligence.

First came AI-generated bands like The Velvet Sundown, trying to pass themselves off as original and authentic music, despite being generated out of existing references. Now, AI-generated songs are sneaking in the profiles of human artists, more or less known, like Erreway, and often deceased, like Angela Lansbury.

From MySpace to AI: Who gets to generate art?

If everyone can generate something that looks like art, will we continue to value the people who spent years learning how to make it?

In the late 2000s and early 2010s, technologies to produce art digitally became increasingly affordable, while new platforms made it possible for creative work to be distributed and discovered online without relying on the traditional gatekeepers of the arts and entertainment industry.

It was the era of bedroom musicians getting discovered after posting songs on MySpace, independent filmmakers building audiences by publishing their work on YouTube and Vimeo, and photographers, illustrators and designers developing their craft while sharing it through Instagram.

Having more democratic access to creative tools did not replace creativity. It expanded who could participate in it. The barriers to creating became dramatically lower, but creating still required skill and knowledge. The tools still had to be mastered. Behind every artist were countless hours spent learning a craft before anyone was paying attention.

Fast forward to 2026, and anyone with a subscription to ChatGPT or another generative AI platform can generate something that resembles a song, an illustration, a screenplay or even a video in a few minutes.

Notice that I said generate, not create.

I’m in a toxic relationship with my algorithm

Algorithms influence what we see, read, and watch online. Knowing how they work can help us use them more intentionally.

Digital algorithms were meant to improve my experience online, giving me a more tailored feed of content (and ads) that matched my needs.

I was not supposed to notice how precisely they were suggesting music that matched my exact mood, videos that appeared right when I was thinking about them, or shows and movies I had not searched for but somehow already wanted to binge.

At first, it felt like a win-win situation.

The platform got more engagement from me. I got a more personalized experience. The internet felt less chaotic and more intuitive.

Turns out, it was not that simple.

What happens to the videos we record?

The evolution of domestic video into short-form vertical content is reshaping how we document, tell, and remember stories.

Domestic videos have gone through a quiet but significant transformation on recent decades, both technological and social.

Not so long ago, these recordings of celebrations and family trips lived only in private collections, recorded on analog formats that were rarely revisited and built up dust on shelves.

That started to change with the technologies to upload and watch video online, particularly with the foundation of YouTube in 2005.

Fast forward to a couple decades later and now domestic videos are produced and shared instantly across platforms. We have slowly got used to watching streams of content from friends and family, but also from strangers.

Recording, sharing and watching domestic videos is now constant, embedded in daily life, and often public by default.

AI changes referral traffic

Artificial intelligence alters content discoverability across platforms, redefining what distribution strategies should look for.

The sources that defined referral traffic online used to be clearer just a few months ago. Search brought people looking for something specific, either evergreen or breaking; successful social media posts could determine clear boosts in page views and direct traffic was a sign of brand recognition, even better when it was regular and loyal visitors.

Artificial intelligence and other algorithmic changes from tech companies have messed with the way we discover and consume content, with some sources of external traffic heavily hit while others have managed to survive or even thrive.

People are still consuming content, probably more than ever, but the way they get to it is less linear and, in many cases, harder to track.

As the “Ask Google” habit shapeshifts to “Ask the Chat” (ChatGPT), how much has AI changed how we distribute and discover content online?

Videos that remain useful last longer

Strategies for videos that remain relevant, searchable, and impactful, driving audience engagement and sustained newsroom success.

In digital news media, success is often defined by the newsrooms that published accurate information first. Platforms also reward immediacy and audiences scroll through an endless stream of updates that can feel obsolete within days.

But a more durable truth persists: Evergreen content grows over time, answers relevant questions, surfaces in search results, supports ongoing reporting and strengthens a newsroom’s credibility long after the initial spike in traffic fades after breaking news.

For video, that idea is a structural choice that shapes how stories are conceived, reported, edited and distributed.

Adapting ideas for multi‑platform storytelling

Share stories effectively across web, social, video, and newsletters to connect with audiences on every platform.

In today’s digital landscape, audiences encounter stories and information in dozens of places: websites, YouTube channels, social media accounts, email newsletters, messaging apps, and each comes with different expectations, attention spans, and ways of engaging.

Adapting ideas for multi‑platform storytelling means making editorial choices that preserve your original intent while reshaping tone, structure, and delivery to match the platforms where your audience lives and interacts.

This article offers an editorial perspective with practical approaches that extend beyond simple repurposing.

Video consumption patterns reshape editorial decisions

Algorithmic feeds and mobile-first vertical video are transforming workflows and the way news and information reach audiences.

Video has lived on social platforms for decades now, but the dominant mode of consumption has shifted in recent years.

Discovery increasingly happens through algorithmic feeds rather than search or deliberate selection.

At the same time, vertical formats have become the default, rather than the exception, as mobile consumption takes the lead.

These changes are subtle but consequential.

They influence not only how video looks, but also how editorial decisions are made and how production workflows are implemented.

How to design content around audience needs

Designing content around audience needs helps you build loyalty, boost engagement, and create stories that perform across platforms.

Designing content around audience needs is one of the most reliable ways to create stories that work across platforms, build loyalty and reach people who may not yet be familiar with your work.

This goes beyond questions about format or platform.

It asks content producers to understand what people are trying to accomplish when they open an article, watch a video or search for context on a topic that affects their lives.

When you identify those needs early, every editorial and production choice has a purpose and you gain a clearer sense of what the audience is looking for, what the story must deliver and how it should live across platforms.