We have a bigger-than-usual edition, as we have more than usual to share. There are a lot of audio clips this week, all of which will take you to the main post on the website. If that’s not your thing, scroll down for new artists fresh from the RADAR.

It wasn’t that long ago that we made AI music AND published it. We wanted to see just how easy it is to get onto streaming platforms. Turns out very easy and very cheap.

We’ve been tracking it ever since, and it still gets the occasional play. It was an interesting process, but now the platform we used has been upgraded with the ability to add your own voice, which can be highly manipulated and put on instrumentals.

We’ve seen an increase in this type of submission lately, so it’s a feature a lot of people are using. It got us thinking - how easy is it?

We also tested Gemini Lyria 3.

Starting with the worst thing we created this week, here is God of the Algae - we highly recommend you turn it down. This was a totally hands-off experiment; all we did was use the Bad Music prompt.

Here are the lyrics, also ‘written’ by Gemini:

Cool, our favorite bit is the slightly off-key humming used as lyrics rather than just a direction. Interestingly, everything that Lyria 3 produces has a built-in SynthID Watermark, so it will always be identified as AI. But since the output is arguably terrible, we don’t think anyone would think people made anything it generates at the moment.

Onwards, to Suno then. Purely for our readers, and something we won’t be publishing on any streaming services, we decided to play around with spoken word - because we can’t sing. Recording your voice is easy (we’re using Suno + Rode NT mic).

All you need is a reasonable microphone and about ten minutes.

If you give it no direction and add no lyrics after recording your vocals, you’ll get something that has very few actual words. It will take what you recorded and smash it together into something that gets worse and worse the more you listen.

Because why not trees the power in every bear? Lyrical genius.

Interesting, we had assumed that what we recorded would be what was used in the track we were making. But it didn’t; it needs lyrics added during the creation process. It also added an accent, which seems to be a common theme with anything that clones your voice. And, depending on the style of the instrumental, the accent it applies changes…

So what can you do with it that sucks less?

Fresh from listening to Choose Life from PF Project ft. Ewan McGregor (Trainspotting)…

Instead of choosing life, a compact disc player, and electrical tin openers (snippet of original lyrics for you there), we chose convenience and a Techno / Acid trance instrumental. We kept the super cynical lyrical styling.

And this is what we got in the end.

Far from genius. What is interesting is what it does with your recorded vocals. This time, it hasn’t added the accent; instead, it is much closer to reality, and since we input the lyrics, it has real words and sentences, so that’s great. Despite the lyrics and style being dangerously close to the original, there were no copyright warnings. However, when we tried it with something close to Adele’s Someone Like You it did flag.

Spoken-word cadences are harder for auto-copyright to flag, unlike more recognizable melodic vocals.

But what if we used the same spoken word sample we gave it to make a folk song? How bad could it be? Well, it did indeed make us sing.

Voice manipulation has been around for a long time, like auto-tune, for example. Most of the biggest acts in the world will use auto-tune to help them create the music that they want. We’ve known that for years, and for the most part, no one cares.

This feels like a huge step forward. The difference is that when real singers perform live, they can actually sing. Now we will have a slew of people, like us, who can’t sing at all, but Suno made it happen.

No matter how big a fanbase you gather, you’ll never be able to perform it live - so the question is, what is the point? Some artists have hardly ever performed live, Enya being one. There are some rare performances, but never a tour. But not everyone can be Enya.

Or can they?

In March, a singer-songwriter, Benedict Cork, posted an unfinished snippet of a song on TikTok. Only days later, he was getting congratulated on his release. The only problem was that he didn’t release it. Someone had taken the start of the track and constructed the rest of it using AI and published it on Spotify.

It hits dangerous territory because it means that real artists will start to reconsider sharing their works in progress, and have to pin all hopes on something already released. It makes it difficult for them to promote their work and start that anticipation, and get the pre-saves that benefit the artist.

So we guess, by that measure, you could potentially take an Enya track, chop it down, and rebuild it.

Record labels and marketing leads at the end of 2025 predicted that, as an audience, we’d seen more BTS and teasers. While so far that has been the case, with the new AI capabilities, will it force the creative process back behind closed doors?

Who knows? We’re in new territory.

A couple of weeks ago, we launched RADAR. It has become our spot for sharing artists that you probably haven’t heard of. Here are a couple of our favorite finds. Rather than pack more links in here, you can go searching on your music streaming platform for what takes your fancy.

Shanghai Treason. A band you can see live, which is cool. We fell hard for a track called Drowning Heart, it’s celtic, banjo, punky goodness.

Hang Linton. In terms of genre, we struggled to find the right one. If you like Scroobius Pip, you’ll enjoy SALE. And if you haven’t heard of Scroobius Pip, then you’re welcome.

Sophie Hunter. ALT+POP+RAP. Not for the faint of heart or easily offended, we’re not going to recommend our favorite song; instead, we’ll say it’s the song at the top of her Spotify profile. With Walk being a close second.

What else is on repeat this week? Dominic Fike - Babydoll, Michael Jackson - You Rock My World, Dermot Kennedy - The Weight of the Woods (whole album), Arlo Parks - Ambiguous Desire (whole album), and Sneaker Pimps - Spin Spin Sugar are in heavy rotation.

The website had a big update. You’ll find a ticker tape of music news, two ways to find music, one new artist randomizer, and one for the weird songs for our more obscure playlist. The RADAR page now has a way to submit music to us and add your music straight to a playlist.