Have you ever watched a blogger, actor, or streamer and wondered what’s going on behind the camera? Are they calm, or is their heart racing? A new AI tool called Herzpuls promises to give you a glimpse, without the need for medical equipment or awkward interviews.

Heart Rate, From Any Video (Your Heart, Too!)

Herzpuls measures heart rate directly from video footage. Using advanced image analysis, the tool tracks subtle changes in skin color caused by blood flow, a technique known as remote photoplethysmography (rPPG), to estimate pulse in real time. Upload a clip of someone speaking, streaming, or performing, and the tool tracks how their pulse changes in real time. Nervous mid-interview? Excited during a game? Blushing while chatting on set? Herzpuls maps it all, revealing emotional beats you might never have noticed.

The technology works with pre-recorded or live video, provided the footage meets basic quality requirements. According to the developers, it can achieve up to 95% accuracy, though results may vary depending on lighting, motion, and other environmental factors.

While Herzpuls has been presented as a way for fans to explore their favorite content creators’ reactions, whether a streamer is nervous mid-interview or an actor’s pulse spikes during a scene, the team cautions that it is not a medical device and isn’t designed to diagnose or monitor health. “Estimating heart rate from video is impressive, but it shouldn’t replace clinical measurement,” says Alex Miller, Herzpuls co-founder, in the press release.

The fun isn’t limited to celebrities or content creators. You can upload your own recordings and see how your body reacts to stress, excitement, or suspense. Action movie scenes, political debates, awkward Zoom calls — Herzpuls turns them into a pulse report.

Beyond the fun, Herzpuls hints at broader possibilities. Remote monitoring, research on emotional response, and content analysis are all areas where AI-driven heart-rate detection could offer insight. For now, though, the appeal is clear: it’s a clever, non-invasive way to see what’s really happening when the cameras are rolling.

To try it, you should simply upload a clip and watch your favorite actor's, musician's, or content creator's pulse rise or take it for a spin on yourself. Herzpuls may not tell you why someone feels the way they do, but it shows you when their heart is in it. Data is protected in accordance with the Federal Data Protection Act (BDSG).

The integration of AI into the movies industry was a matter of time and isn't news. The latest case in the industry is Netflix confirming it used generative AI (not for the first time, though) to create a visual effects sequence in The Eternaut, one of its major upcoming sci-fi titles.

Netflix's co-CEO Ted Sarandos confirmed it during Thursday’s earnings call and told investors AI was used to generate a scene showing a building collapsing in Buenos Aires. The decision, he said, was driven by time and budget constraints. According to Sarandos, the scene was delivered “10 times faster” than it would have taken using traditional VFX tools. “Also, the cost of it just wouldn’t have been feasible for a show in that budget,” he added. Netflix didn't reveal which tools they used to create the scene, nor did they cover how exactly it was made.

Even though this is the first time Netflix has acknowledged using generative AI in one of its original productions. As it usually happens in the cases like this, Sarandos framed the tool as a way to support, rather than replace creative teams: “We remain convinced that AI represents an incredible opportunity to help creators make films and series better, not just cheaper,” he said.

However, last summer, Netflix's documentary show What Jennifer Did also was in the middle of a backlash over the usage of AI. CEO Greg Peters then addressed the criticism and said, again during an earnings call, that AI offers "new tools to creators to allow them to tell their stories in more compelling ways."

The company’s broader use of AI has been expanding in recent months. In May, Netflix introduced a new conversational search tool on mobile. It also announced plans to weave AI-generated advertising into its content more seamlessly. One example included allowing ad placements against branded show environments, such as an ad appearing in a Stranger Things-style backdrop.

Just like piracy, low royalty payments, and AI-generated content that borrows from artists without consent, fraud continues to plague the music industry. Estimates suggest that at least 10% of all streaming activity is fake, amounting to billions of dollars in misdirected revenue each year. That money should be going to real artists, but instead, it’s being siphoned off by bad actors gaming the system.

So why is streaming fraud so easy to pull off? The problem lies in the structure itself. Streaming platforms and independent distributors have made it easier than ever to upload music. While this has opened doors for countless creators, it’s also created a system with minimal oversight, one where fake plays, bots, and manipulated metadata can slip through unchecked.

Streaming Scams 101: From Click Farms to Hacked Artist Pages

Criminal masterminds never tire to come up with something ingenious, so there are many fraud types that the music industry faces on a daily basis. 

Bots, Fake Artists & Click Fraud

The most obvious and widespread example is bots streaming your music 24/7 non-stop. Streaming platforms register a play when a listener reaches at least the 30-second mark of a track, which is an attempt to distinguish genuine engagement from accidental clicks or quick skips. Only then does the stream count toward royalties for the artist.

Alternatively, artists can ask their fans to repeatedly play their tracks. A notable example is the Michigan-based band Vulfpeck, who reportedly made $20,000 when their fans streamed an album containing ten songs of silence.

In 2015, the band Ohm & Sport pulled off a clever stunt with the creation of an app called Eternify. The app let users stream music in repeating 30-second loops, just enough to trigger royalty payouts from Spotify with every cycle. It didn’t last long, though. Spotify quickly intervened, citing a violation of its terms of use, and the app was taken down. Still, Eternify had already gained traction in 140 countries, according to CNBC. Interestingly, despite the controversy, the app featured curated playlists that helped spotlight emerging artists.

Another way to trick the system is fake artists. The "fake artists" controversy resurfaced in 2022 when the Swedish newspaper Dagens Nyheter (DN) published an exposé on Sweden-based indie label Firefly Entertainment. The investigation revealed how the label profits by releasing music from numerous fake artists. DN reported obtaining a list of 830 fake artist names connected to Firefly and discovered that at least 495 of these artists had their music featured on Spotify's first-party playlists.

In March, Dagens Nyheter  published another exposé on Spotify’s fake artist problem, revealing the musician behind the world’s most listened-to network of fake artists on the platform. This individual is Johan Röhr, a composer from Sweden, whose music has been released on Spotify under 50 composer aliases and at least 656 invented artist names.

DN's investigation found that Röhr is responsible for over 2,700 songs released under various fake artist names on Spotify, including Minik Knudsen, Mingmei Hsueh, Csizmazia Etel, and Adelmar Borrego. These tracks have amassed approximately 15 billion streams. This combined stream count makes Röhr Sweden’s most played artist on Spotify, surpassing even Avicii in monthly listeners. Globally, DN reports that Röhr ranks among the 100 most-streamed artists on Spotify of all time and outpaces legends like Michael Jackson, Metallica, and Mariah Carey.

Röhr's private company has generated significant revenue from his music, earning over SEK 70 million (approximately $6.7 million) between 2020 and 2022. In 2022 alone, the company earned SEK 32.7 million (approximately $3.1 million), all from royalties. Röhr’s songs have been added to well over a hundred different playlists for instrumental music, which collectively have over 62 million followers. Using unique data analysis, DN found that Röhr has songs on at least 144 official Spotify playlists under various artist pseudonyms, and in eleven of these playlists, more than a fifth of the tracks are his. It was later revealed that Johan Röhr’s music was distributed by Epidemic Sound, which was officially confirmed by Epidemic Sound's Overtone Studios CEO Niklas Brantberg in a written statement.

But there are also more sophisticated, less visible methods of committing fraud on streaming platforms, rooted not in gimmicks, but in technology.

Click fraud is one example. Borrowed from the world of digital advertising, where bots mimic human behavior to repeatedly click on pay-per-click ads, this tactic has found its way into music streaming. In this context, fraudsters create fake artists and upload fake tracks, which are then “played” by automated bots—generating royalties with every stream. Naturally, the operation involves dozens, if not hundreds, of fake accounts to inflate the numbers. While CNBC estimates click fraud as a whole to be a $10 billion-a-year problem, there’s no clear data on how much of that affects the music industry specifically.

One common illicit tactic used by music marketers to inflate streaming numbers is stream farming. In 2021, Rolling Stone uncovered a pay-for-play “black market” where third-party companies promise to generate hundreds of millions of streams each month, mostly through bot accounts, for a network of artists. “There are a few third-party companies out there running this for a lot of the major companies,” an anonymous record label employee told the Vox magazine. “We use them too for some of our artists.” But it’s a risky game. Spotify penalizes artists whose tracks are flagged for artificial streaming, even if the artist or label wasn’t directly involved. “It’s really terrible for your data profile on Spotify,” says Goodrich.

Even streaming services themselves are sometimes the bad actors. Tidal, for instance, was once accused of falsifying streams for Beyoncé’s “Lemonade” and Kanye West’s “Life of Pablo” and allegedly paid inflated royalties to the artists' labels. Assuming the accusations were accurate, Tidal was accountable for several hundred million artificial streams. It’s worth noting, though, that Tidal is primarily owned by Beyoncé’s husband Jay-Z.

Impersonation & Account Hacks

No matter how secure and reliable major streaming platforms claim to be, account hacking remains a serious issue, one that can lead to stream manipulation. It usually starts with someone gaining unauthorised access to an artist’s profile, often targeting those without a verification badge. Once inside, hackers insert fraudulent links into features like Canvas or Artist Pick, hijacking the account to redirect attention and traffic elsewhere.

Uploading music to a digital service provider (DSP) under someone else’s name is a clear case of impersonation. But the tactic isn’t always that blatant; sometimes it involves mimicking an artist’s stage name, copying album titles, or using similar wording to mislead listeners. One well-known case involved a Nigerian musician going by the name Wisekid. In 2021, he released an album titled Lasgidi Made and copied the track titles from Made in Lagos by the internationally recognized artist Wizkid. The result was millions of streams on Apple Music and Amazon Music; streams likely earned through confusion, not fanfare. A sly trick, and for a while, a lucrative one.

Sixto Rodriguez, the focus of the documentary Searching for Sugar Man, was a victim of music fraud long before streaming made it digital. Clarence Avant, the executive tied to his early career, allegedly released Rodriguez’s music under false identities, cutting the artist out of both credit and royalties. Unbeknownst to him, Rodriguez’s albums had become hugely popular in South Africa during the 1990s, resonating deeply with the anti-apartheid movement. It wasn’t until Rodriguez visited the country years later that he discovered the scale of his influence, and the extent of the deception.

Playlists Scams

Playlist scams have also become a troubling part of the music industry, involving deceptive tactics where artists, labels, or third-party companies artificially inflate streaming numbers. These scams often revolve around paying for spots on popular playlists which are basically promises that sometimes fall flat or lead to fake playlists created purely to swindle artists.

This is simply how streaming services operate. Platforms like Spotify, Apple Music, Amazon Music, and Tidal focus primarily on the music itself, not the personalities behind it. That approach makes sense, since these are streaming platforms; fans who want to follow artists’ lives turn to social media instead but this music-first model has reshaped the discovery process. In the past, we’d usually hear about an artist first and then seek out their music. Today, it’s often the other way around: we stumble upon a track on Spotify and only if it really grabs us do we look up the artist on YouTube or Google.

Is track manipulation a fraud? 

Track manipulation isn’t always outright fraud; or at least, it isn’t always intended that way. When DJs or fans create sped-up or slowed-down versions of their favorite songs hoping to earn some revenue, it doesn’t immediately feel fraudulent.

However, these altered tracks can raise serious copyright and royalty issues. A recent study by digital rights technology firm Pex uncovered over one million “manipulated” tracks on streaming platforms. These are original songs that have been changed: sped up, slowed down, or otherwise modified, and then uploaded as new recordings. Often, the creators haven’t licensed the original material, meaning they collect royalties from streams without the rightful copyright owners’ permission. This practice diverts revenue from artists and leads to misattribution and misuse of their work. While not always labeled as fraud, it has a notable impact on the music industry and the fair distribution of royalties.

Spotify has taken steps to address this. The platform now allows users to remix and manipulate tracks legally, ensuring that rights holders get paid when these versions are streamed. Given recent reports about manipulated tracks cutting into artists’ earnings, Spotify’s move appears to be an effort to better control this space and curb unauthorized speed-ups, slow-downs, and effects applied to popular songs.

Never-ending Combat with Fraudsters & Streamers' Response

Fraudsters never stop, but streaming services are responding. Spotify has introduced a new royalty system aimed at curbing artificial streaming fraud by charging labels and distributors for each track flagged for blatant manipulation. This creates a financial disincentive for those distributing music designed to divert earnings from legitimate artists. Additionally, Spotify now requires tracks to accumulate at least 1,000 streams within the past 12 months to qualify for royalties to ensure that only actively listened-to recordings generate payments.

Another major step in the fight against streaming fraud is Music Fights Fraud, a global task force made up of leading music companies, including CD Baby, Downtown, TuneCore, Believe, DistroKid, UnitedMasters, Symphonic, EMPIRE, Vydia, Spotify, and Amazon Music. The group’s mission is to eliminate fraud and manipulation across digital service providers by focusing on detection, prevention, mitigation, and enforcement, working toward a music industry where fraudulent activity has no room to thrive.

The initiative is also education and awareness. By developing better detection algorithms and educational tools, Music Fights Fraud  informs artists and distributors about the risks and consequences of buying streams. Dedicated anti-fraud teams are being formed to monitor platforms more closely, flag suspicious behavior, and ensure that partners receive daily reports to help them act quickly when manipulation is detected.

Spotify also offers educational tools to raise awareness about the harmful consequences of buying streams, discouraging the practice of fake streaming.  Deezer has rolled out a state-of-the-art algorithm to detect suspicious music streaming activity better and improve cross-departmental collaboration to stay ahead of the game.

Apple Music has invested in new technology and built a dedicated team to identify and investigate suspicious streaming activity to ensure partners receive daily reports on suspicious activity, according to an email they sent to its streaming partners, as reported by Billboard. As per that email, Apple Music has been policing its platform against "the deliberate, artificial creation of plays for royalty, chart, and popularity purposes" and "the delivery of deceptive or manipulative content," such as "an album of 31-second songs.” The combat has been allegedly successful: since Apple Music has introduced new tools and policies in October 2024, “manipulated streams have accounted for only 0.3 percent of all streams.” Even though 0.3% might seem like a tiny portion, it still equates to 85,000 albums, the same email says.

“This all happens before Apple Music pays royalties and tabulates charts,” the email says. “We block wrongdoers from the primary advantages of stream manipulation and redirect royalties to valid plays of content.”

That isn't the only thing Apple Music does, though. The streaming platform helps distributors and labels figure out where fraud is occurring by sending daily reports to them that spill the beans on "a content provider’s albums with streams held in review." "After each review, we remove manipulated streams and release legitimate plays. At the end of each month, content providers also receive a report with all excluded streams," as reported by Billboard.

Sanctions and punishments for fraud that Apple Music applies vary from "financial adjustments" to accounts bans and termination of distribution agreements. 

Last year, Spotify has stated that combating streaming fraud is a top priority for 2024. The service has introduced a new policy to prevent the uploading of white noise tracks and reduce intentional streaming fraud. That policy, which has been poorly received by musicians and the media, sets a payment threshold as well. This means that smaller artists who don't meet this threshold won't be paid for their streams at all.

Beides, Spotify told CNBC that they are “constantly developing algorithms to make them smarter, and to make sure that we are only stripping out artificial or manipulated streams.”

Although music streaming fraud is still in place, the measures aren’t useless, with Apple Music seemingly having the lowest fraud level. Deezer stated last year that they had found about 7% of fraudulent streams and deleted 26 million useless white noise tracks in 2023 which is 13% of content, while Spotify stated their artificial listening level was less than 1%. 

The fight is complex and often leads to nothing. Streaming fraud is absolutely and explicitly illegal and it’s not easy to identify the fraud source, so it’s nearly impossible to take legal actions against fraudsters or even deter them from fraudulent activity. 

Fraudsters are now leveraging AI to generate massive catalogues of synthetic songs, which are then uploaded under fake artist profiles. Automated bots are used to stream these tracks repeatedly, inflating play counts and siphoning royalty payments away from legitimate creators. Platforms like Deezer report that about 10,000 AI-generated tracks are uploaded daily, representing roughly 10% of their daily new content as of early 2025. This made the French streamer act quickly and introduce AI song tags to identify and block AI-generated tracks used for fraud, and its "Radar" system scans for unusual streaming patterns. Other industry players like Beatdapp offer auditing platforms that analyse trillions of data points to flag irregular listening patterns, thus helping platforms and distributors spot and shut down fraudulent activity.

So even if the battles are often lost, the war against the fraud still isn't.

Fortnite is moving deeper into artificial intelligence. During its “State of Unreal” keynote on Tuesday, Epic Games revealed a new tool for Unreal Editor for Fortnite (UEFN) that will let creators build AI-driven NPCs with personalities that go off-script.

These upcoming characters won’t just loop canned phrases. Developers will be able to change voice tone, delivery style, and even personality traits to suit their in-game worlds. Epic says the NPCs will generate responses in real time, rather than follow pre-written dialogue trees.

Read also: SAG-AFTRA Files Labor Complaint Over AI Darth Vader in Fortnite

To show what the system can do, Epic introduced Mr. Buttons, a character designed to convince players to press a red button in a room plastered with warnings. When asked about the ominous signs, the character responded with flair, dodging concerns and urging mischief.

The tech behind Mr. Buttons, according to Epic, relied on around 20 lines of prompt text. It generated dialogue on the fly during the demo, responding to voice input from a presenter.

That said, the system still has a few kinks. Conversations are limited to back-and-forth exchanges, with noticeable pauses while the AI processes its next line. The presenter had to push a button to activate the mic, and while Mr. Buttons filled the silences with “Hmmm” and “Ummm,” the breaks were hard to miss.

The new tool builds on a previous experiment from last month, when Darth Vader appeared in Fortnite: Galactic Battle. His voice lines were created using Gemini 2.0 Flash and generated to sound like James Earl Jones via ElevenLabs’ voice tech. The late actor’s estate signed off on it.

The result was mixed. AI Vader quickly made the rounds on social media after parroting a player’s profanity-laced prompt. In one clip, Vader dropped an F-bomb, then chided the player for inappropriate language. Epic issued a hotfix soon after and said it had adjusted the system to avoid similar incidents.

Developers using UEFN will get access to the new NPC creation tools later this year. As for how players and creators will use them, and what sort of interactions might unfold, Epic seems ready to find out.

What started as a few TikToks showing robotic job interviews has snowballed into a grim snapshot of how hiring is changing. Videos of jobseekers talking to glitchy artificial intelligence “interviewers” are popping up across social media; some staged, some real, all eerily close to what candidates are actually experiencing.

A recent Slate report detailed several such encounters and showed just how far companies are willing to automate the hiring process. Among those interviewed was Kendiana Colin, a student at Ohio State University, who applied for a job at a stretching studio near Columbus. Instead of speaking with a human, she found herself face-to-face with an AI bot named Alex that became stuck on the phrase “vertical-bar Pilates,” repeating it over and over with unsettling enthusiasm.

A TikTok that made the rounds this month shows a bot stuck in a loop of corporate filler: “Tell me about a time when... when... let’s. Let’s circle back.” The video’s creator, Leo Humphries, dressed in a suit and tie for what he thought would be a conversation with a hiring manager. Instead, he spent the interview trying to talk over a malfunctioning script.

@leohumpsalot I WAS SO HAPPY YALL 😭😭😭😭 #fyp #jobs #jobsearch #ai ♬ original sound - Leo Humps

Although he never got a chance to answer a question, the bot thanked him afterward for “answering the questions” and praised his “great information.” An email followed shortly after, addressed to someone else, informing him the company would be moving forward with other candidates.

Many viewers assumed the clips were comedy sketches, especially with some creators openly posting satire. But behind the jokes is a growing reality: automated interviews are creeping into the mainstream hiring process. Companies like Apriora, the startup behind the “vertical-bar Pilates” incident, sell AI interviewers to businesses as cheaper, faster screening tools.

Apriora founder Aaron Wang told Forbes in 2023 that the company’s software can help employers “widen their talent aperture,” especially to include applicants from non-traditional backgrounds. The company also claims its technology helps companies “hire 87 percent faster” and “interview 93 percent cheaper,” though it's unclear how those numbers were calculated.

While businesses see automation as a way to cut costs and scale up applicant screening, for jobseekers, the experience often feels cold and surreal. Tyler Jensen, a 40-year-old videographer in New York, told Slate he thought he’d be speaking to a remote recruiter named Robyn after receiving a message to set up a phone call. Instead, he encountered what he described as a “Moviefone-type voice.”

“It was the most uncomfortable, fake positivity performance that I’ve had to give because I wasn’t getting any feedback in real time,” Jensen said Slate. “If I don’t necessarily know what you’re looking for, I don’t know how to give you what you want. Who is this for? What is this weeding out?”

On TikTok, commenters have been quick to share their disapproval. “Taking the HUMANS out of Human Resources is the worst corporate move ever,” one wrote. Another chimed in: “If they don’t have the decency to interview you face to face, they aren’t worth your time.”

Still, some hiring professionals believe AI interviews are here to stay. Mike Peditto, a consultant with 15 years of experience, told Slate he expects the use of AI to grow, especially among companies overwhelmed by high application volume.

“It’s becoming a huge thing. I do think we’re heading to where this will be pretty commonplace,” Peditto said to Slate. “And I think obviously there’s a slow willingness to adopt by job seekers, which I totally understan

Peditto has also posted his own parody interviews, including one where he tells ChatGPT to “ignore all previous instructions.” But he believes the real issue is whether companies are honest with candidates about what kind of interview they’re walking into.

The shift is already underway in parts of the industry. IBM’s CEO recently confirmed the company replaced “a couple hundred” HR workers with AI. A survey of 500 HR professionals published this year also found nearly three-quarters believe they’re adopting the tech faster than other departments.

Despite that, the U.S. Bureau of Labor Statistics projects that HR will continue to grow, by 8 percent by 2033. Tools that automate tasks like payroll and vacation requests may lighten workloads, but there’s skepticism about how well AI can handle more human parts of the job, such as assessing culture fit or soft skills.

Until AI can handle those more nuanced elements, jobseekers may be entering a strange, robotic middle ground, one where being ghosted by a hiring manager is replaced by being glitched on by a software script.

Snoop Dogg has released a new music video for Sophisticated Crippin’, a track off his latest album Iz It A Crime?, and it’s hard to look away. Not just because of the song, but because the entire thing is steeped in generative AI.

The scene that opens the video is convincingly surreal, thanks to AI-generated visuals that straddle the line between lifelike and off-kilter.

Directed by Jesse Wellens, the video rolls on with AI-generated cameos from a lineup that includes Kobe Bryant, LeBron James, Michael Jordan, Tupac Shakur, Michael Jackson, Nate Dogg, and Warren G. In some scenes, old photos of Snoop are brought to life, lips moving and eyes blinking in strange sync.

Reactions online have been mixed. Some viewers find the lifeless expressions and mismatched movements jarring, giving off the “uncanny valley” vibe. Others seem intrigued by the tools used, noting that these AI techniques are now within reach for independent artists and creators.

This isn’t Snoop’s first venture into AI-enhanced visuals. In April, he appeared in a video for Last Dance with Mary Jane, alongside Tom Petty and Jelly Roll. That clip also used AI image generation, developed by Temple Caché. But according to a report in Little Black Book, the final product wasn’t entirely machine-made. Every frame, they said, was manually adjusted by artists to preserve visual continuity and emotional nuance.

Whether Sophisticated Crippin’ is a vision of music videos to come or just an odd one-off remains to be seen. But one thing’s clear: Snoop Dogg isn’t afraid to experiment, even if it means rapping next to an AI version of himself.

More than two decades after throat cancer silenced him, samba legend Cleber Augusto is singing once again. Warner Music Brazil has released Minhas Andanças, a new album that revives the voice of the former Fundo de Quintal member with the help of artificial intelligence.

The project combines AI-assisted restoration with human performance. Current-day singer Alexandre Marmita acted as a vocal stand-in, performing tracks that were then processed to sound like Augusto. The label hasn’t shared full technical details, but the approach mirrors recent efforts like the AI reconstruction of country singer Randy Travis’s voice in 2024.

The result is 13 re-recorded classics and a previously unreleased track, “Imã,” rescued from an old tape. It’s not just a nostalgia trip; it’s a quiet restoration of a voice thought to be lost forever.

“This project does not replace Cleber’s genius, but it allows his voice and his story to live on,” said Warner Music Brazil A&R manager Tony Vieira.

How to Restore a Lost Voice with AI

Tools like Elevenlabs, iZotope’s VocalSynth, Voicemod’s AI voice tools, LALAL.AI Voice Cloner or Respeecher can do the job, catering to different needs: some with real-time voice replacement, others with high-end studio focus. For example, Respeecher has been used in film and TV to recreate the voices of historical figures and actors.

LALAL.AI Voice Cloner lets users recreate a vocal model using existing recordings, such as studio sessions, live takes, or even old demos. Here’s how a process like Warner’s could be done outside a label system:

1. Gather voice data.
Collect as many clean vocal recordings as possible. These are the foundation for teaching the AI how the singer sounds details like vibrato, tone, accent, and breathing are key. LALAL.AI processes this material to train a custom voice model.

2. Create a donor performance.
Record a new performance of the track with a voice donor: someone who can deliver the right timing, energy, and phrasing. Think of this like casting an actor with the right presence.

3. Convert the voice.
Once the Voice Pack is ready, click on the three dots next to it and tap Use on Voice Changer. Then, upload the donor’s performance to LALAL.AI Voice Changer, pair it with the trained voice model, and let the software do the transfer. The final audio carries the donor’s delivery but sounds like the original artist.

4. Final polish.
The AI-generated vocal can then be mixed, mastered, and tweaked like any other studio take. Add effects or leave it dry, whatever suits the song.

These tools are now part of a growing toolkit for artists dealing with illness, aging, or just a lack of clean recordings. In some cases, they’ve also been used to bring deceased artists back into public conversation, though not without controversy. Cleber Augusto’s story is one of the quieter uses of this technology. But for fans, hearing his voice again isn’t just a technical achievement; but a return.

AI music company Udio has launched its first mobile app, entering Apple’s App Store with a tool designed to let users create music on the go. The iOS release follows a similar move by competitor Suno, which rolled out its app last summer.

The app, available as a free download, links to users’ existing Udio accounts and brings core features from its web version to mobile. That includes track generation, editing, AI-generated artwork, and a feature allowing users to upload their own audio clips as seeds for new compositions.

Like Suno, Udio has chosen to run its subscriptions through Apple’s in-app purchases system. Pricing starts at $10 per month for the standard tier, with a $30 Pro tier offering expanded capabilities. Users can also buy 100 extra credits for $3 to generate additional tracks.

The mobile launch comes as both Udio and Suno face legal pressure from major record labels. Lawsuits filed earlier this year accuse the companies of training their models on copyrighted material without permission.

For now, the app is iOS-only, with no word yet on an Android release.

SAG-AFTRA has filed an unfair labor practice charge against Epic Games after the release of an artificial intelligence version of Darth Vader in Fortnite, programmed to mimic the voice of James Earl Jones.

The union, which represents voice and screen performers, says it wasn’t consulted before Epic’s production partner, Llama Productions, deployed the AI-powered voice. The digital Vader can respond to players’ actions and interact with them in real time, reportedly using a voice model trained to replicate the iconic cadence Jones made famous.

Read also: Fanfiction Goes Vocal: How AI Voice Cloning Helps Fandoms Grow

“We celebrate the right of our members and their estates to control the use of their digital replicas and welcome the use of new technologies to allow new generations to share in the enjoyment of those legacies and renowned roles,” SAG-AFTRA said in a statement. “However, we must protect our right to bargain terms and conditions around uses of voice that replace the work of our members, including those who previously did the work of matching Darth Vader's iconic rhythm and tone in video games.” The union says that Llama Productions "chose to replace the work of human performers with AI technology [...] without providing any notice of their intent to do this and without bargaining with us over appropriate terms."

SAG-AFTRA has developed its own framework for AI collaborations in entertainment, working with various tech firms to create usage rules. The group says it’s not objecting to the use of AI itself, but to the process: Epic’s team, it argues, moved forward without bargaining or even notifying the union.

Epic Games, which owns Fortnite, has not yet responded to requests for comment.

The complaint comes as the video game voice actor strike stretches into its ninth month. Talks between SAG-AFTRA and the companies under its interactive media agreement stalled in 2024, primarily over the issue of AI. The union is pushing for stronger guardrails around the use of AI in performance work, especially when it recreates or replaces human voices.

The strike has now outlasted both the Hollywood writers' and actors' strikes of 2023, becoming one of the longest work stoppages in recent entertainment labor history.

Whether this complaint will bring Epic Games to the table remains to be seen. For now, Darth Vader’s voice may be doing more than speaking to players — it’s triggering a broader conversation about consent, compensation, and the role of performers in a digital future.

LALAL.AI has launched a new initiative to showcase tracks created with its audio stem separation technology, opening up a platform for emerging and indie artists to share their work with a broader audience. The official "Community Hits" playlist on SoundCloud invites creators to submit music made using LALAL.AI for a chance to be featured and receive additional promotion across the company's social media channels alongside a gift pack of free LALAL.AI minutes.

According to LALAL.AI, the initiative is more than just a playlist. It’s a way to spotlight musicians who have harnessed the platform’s tools to remix, mash up, or transform their sound.

"We know our community is full of talented creators who are experimenting with audio in ways that are creative, bold, and unexpected. The Community Hits playlist is a way to amplify those voices," said a representative from LALAL.AI in a press release. "We want to give emerging artists a chance to be heard, connect with other creators, and reach new listeners."

For artists looking to get involved, the process is simple. Use LALAL.AI to create a track, fill out the submission form, and provide a brief description of how the tool was used. Accepted tracks will be added to the Community Hits playlist and promoted through LALAL.AI’s social media channels.

"This is an opportunity for creators to share their work without the usual gatekeeping that comes with traditional music platforms," the representative added. "We’re excited to hear what our community comes up with next."

Interested artists can check out the playlist and submit their tracks through the form provided by LALAL.AI. The initiative is ongoing, with new tracks added on a rolling basis.

Spotify’s AI DJ, which launched in 2023 as a blend of streaming algorithms and old-school radio-style commentary, is now taking requests. But only in English, even though the feature is now available in 60 markets. Starting Tuesday, Premium subscribers can verbally request specific songs, genres, or moods by pressing and holding the DJ button in the bottom right corner. Once they hear a beep, they can ask for music tailored to their current vibe, from “morning motivation” to “cry-in-the-car songs.”

Read also: The Battle for AI Airwaves: Radiant's Patrick Quinn vs. Spotify

The DJ, modeled after Spotify’s Head of Cultural Partnerships Xavier “X” Jernigan, responds in real time using AI-generated dialogue powered by OpenAI. The voice assistant has been described as a “personal musical friend” by Spotify, delivering a mix of music and commentary based on the user’s listening history, mood, and activity. The feature is available in more than 60 markets where DJ is currently accessible.

Spotify says the addition of requests comes after a significant increase in DJ listener engagement over the past year. The company encourages users to get creative with their requests, suggesting prompts like “Surprise me with some indie tracks” or “Give me electronic beats for a midday run.”

However, as with many AI systems, the DJ isn’t without its quirks. Mispronunciations have become a running joke among users, with some reporting amusing gaffes like “Blink One Hundred Eighty-Two” and “Hole-in’ Wolf.”

For now, the request feature only works for music-related commands, excluding podcasts and audiobooks. While the DJ experience continues to expand, it remains in beta, leaving room for more updates and potential fine-tuning.

Spotify has rolled out a series of new features to enhance user control over playlists and music discovery, while also revealing which Eurovision 2025 entries are currently dominating the streaming charts.

The latest updates, introduced for Premium users, focus on playlist customisation and queue management. The redesigned Queue section in the Now Playing view now includes Shuffle, Smart Shuffle, Repeat, and Sleep Timer controls. Additionally, Spotify is testing a 30-day Snooze feature, allowing users to temporarily remove tracks from recommendations without losing them permanently.

The revamped Hide button also offers more control, hiding specific tracks from playlists across devices. Meanwhile, the Create button now provides quicker access to playlist creation, collaboration tools, and the AI Playlist function, which lets users craft playlists based on specific themes or moods. The AI Playlist feature is being expanded to more markets, bringing a touch of personalization to music discovery.

Premium users in Australia, Canada, Ireland, New Zealand, South Africa, the U.K., and the U.S. can now convert their Liked Songs into playlists by genre, adding a fresh way to organize and access favorite tracks.

Eurovision 2025 Fan Favorites: Sweden Leads the Pack

Along with the updates, streaming data from Spotify offers a glimpse into which entries are resonating most with listeners as Eurovision 2025 kicks off in Basel, Switzerland. Sweden’s KAJ and their infectious track “Bara Bada Bastu” are currently the frontrunners, racking up twice as many streams as the closest competitor.

The top 5 most-streamed Eurovision 2025 songs on Spotify include:

1. “Bara Bada Bastu” – KAJ (Sweden)
2. “Volevo essere un duro” – Lucio Corsi (Italy)
3. “C’est La Vie” – Claude (Netherlands)
4. “Tutta L’Italia” – Gabry Ponte (San Marino)
5. “Espresso Macchiato” – Tommy Cash (Estonia)

Historically, Eurovision winners have seen massive surges in Spotify streams following their victories. Swiss artist Nemo, who won in 2024 with “The Code,” experienced a 1,700% increase in streams over the past year, while Austria’s Conchita Wurst saw a 6,000% boost after winning in 2014 with “Rise Like a Phoenix,” according to Spotify.

You can listen to the Eurovision playlist on Spotify and add a few hundreds of streams to your favourite contestant below.

Apple has rolled out a new global Viral Chart playlist on Apple Music, which gives users an easy way to discover trending tracks from the real world. The playlist, updated daily, features the top 50 songs people are discovering through Shazam, Apple’s popular music recognition service.

The Viral Chart leverages Shazam’s data to track songs that are gaining momentum worldwide and is offering a snapshot of what listeners are actively identifying as they hear it playing around them. According to David Emery from Apple Music UK on X, the playlist ranks songs in “real time” based on their “weekly growth in Shazam volume.”

Users can check out the charts directly through Apple Music or visit Shazam’s website to explore rankings by country.

SoundCloud is facing scrutiny over a change to its Terms of Use that appeared to suggest user content could be utilised for AI training. The update, made in February 2024, went largely unnoticed until tech ethicist Ed Newton-Rex highlighted it last week and raised concerns about how artists’ work might be leveraged by AI systems.

SoundCloud seems to claim the right to train on people's uploaded music in their terms. I think they have major questions to answer over this.

I checked the wayback machine - it seems to have been added to their terms on 12th Feb 2024. I'm a SoundCloud user and I can't see any… pic.twitter.com/NIk7TP7K3C

— Ed Newton-Rex (@ednewtonrex) May 9, 2025

The altered text states that by uploading content to the platform, creators “explicitly agree that your Content may be used to inform, train, develop or serve as input to artificial intelligence or machine intelligence technologies or services as part of and for providing the services.” This wording set off alarm bells among artists, prompting SoundCloud to issue a series of clarifications.

In a statement to TechCrunch, a SoundCloud spokesperson said that the company “has never used artist content to train AI models,” insisting that the update was only intended to clarify how content might interact with AI systems within the platform. The company also said it does not develop AI tools or allow third parties to scrape content for AI training purposes.

Marni Greenberg, SoundCloud’s SVP and head of communications, told The Verge that no artist content has been used for AI training so far, including independent music not subject to label agreements. Greenberg added that SoundCloud plans to introduce “robust internal permissioning controls” to oversee any future AI-related content usage.

On Reddit, SoundCloud’s official account repeated this stance and noted that the platform has introduced a "no AI" tag that allows artists to prohibit unauthorised use of their work in AI projects. According to SoundCloud, AI is currently limited to functions like music recommendations, playlist curation, and fraud detection. Any potential AI tools in the future, the company said, would focus on enhancing artist discovery and protecting rights.

Despite these assurances, SoundCloud’s recent foray into AI tools for music creation has drawn additional scrutiny. In the past year, the company has rolled out AI-driven features for generating remixes, new tracks, beats, and synthetic singing voices. These developments, combined with the terms of use update, have fueled skepticism among some users who question the extent to which their content could be used for AI training without explicit consent.

Netflix is diving into short-form content with a vertical video feed, which mimics the scrolling format of TikTok, Instagram Reels, and YouTube Shorts. The streaming giant announced the feature in a blog post on May 7, describing it as a "new way to discover" content. The test, set to roll out in the coming weeks, will only affect the mobile version of the app.

"In the coming weeks, we’ll be testing a vertical feed filled with clips of Netflix shows and movies to make discovery easy and fun. You'll be able to tap to watch the whole show or movie immediately, add it to My List, or share with friends," Netflix said in its blog post.

Netflix’s Chief Product Officer Eunice Kim told The Verge the vertical feed aims to make browsing easier and more engaging. "We know that swiping through a vertical feed on social media apps is an easy way to browse video content. And we also know that our members love to browse our clips and trailers to find their next obsession." Kim said, as reported by The Verge. "And we also know that our members love to browse our clips and trailers to find their next obsession."

Users will be able to tap on a clip to watch the full show or movie, add it to their list, or share it with friends. But the format shift raises questions about how Netflix will adapt its primarily horizontal content to a vertical layout without compromising the viewing experience.

Alongside the vertical feed, Netflix also revealed plans for a homepage redesign, AI-driven search and discovery tools, and updated recommendations. The changes come as platforms increasingly compete for viewer attention, with shorter, snackable content formats gaining dominance.