One Tech Tip: Is that song by AI or a human? Here are some things to check
The fictitious rock group, Velvet Sundown, which comes complete with AI-generated music, lyrics and album art, is stoking debate about how the new technology is blurring the line between the real and synthetic in the music industry, and whether creators should be transparent with their audience.
Computer software is widely used in music production, and artificial intelligence is just the latest tool that disc jockeys, music producers and others have added to their production pipeline. But the rise of AI song generators such as Suno and Udio is set to transform the industry because they allow anyone to create songs with just a few prompts.
While some people do not care whether they're listening to AI-generated music, others might be curious to know.
If you encounter a new song that leaves you wondering whether it's 100% made with AI, there are some methods that could reveal how it was created.
Do a background check
If you're wondering who's behind a song, try some old-fashioned detective work.
The 'most obvious cues' come from 'external factors,' said Manuel Mousallam, head of research and development at streaming service Deezer.
Does the band or artist have social media accounts? Lack of a social presence might indicate there's no one there. If they do exist online, examine the kind of content they post, and how long it goes back.
Is there any sign that the artist or band exists in real life? Are there any upcoming concerts and can you buy a ticket for a gig? Is there footage of past concerts on YouTube? Has an established record label released their singles or albums?
Try going to the source. Song creators often — but not always — publish their generated tunes on the Suno or Udio platforms, where they can be found by other users.
The catch is that you'll have to sign up for an account to get access. Users can look up songs by track name or the creator's handle, and browse genres and playlists. But it can still be difficult to spot a song, especially if you don't know the name of the song or creator.
Song tags
Deezer has been flagging albums containing AI-generated songs, as part of its efforts to be more transparent as it battles streaming fraudsters looking to make quick money through royalty payments.
The Deezer app and website will notify listeners with an on-screen label — 'AI-generated content' — to point out that some tracks on an album were created with song generators.
The company's CEO says the system relies on in-house technology to detect subtle but recognizable patterns found in all audio created by AI song generators. The company hasn't specified how many songs it has tagged since it rolled out the feature in June, but says up to 18% of songs uploaded to its platform each day are AI-generated.
Song scanners
There are a few third-party services available online that promise to determine whether a song is human-made or generated by AI.
I uploaded a few songs I generated to the online detector from IRCAM Amplify, a subsidiary of French music and sound research institute IRCAM. It said the probability that they were AI-generated ranged from 81.8% to 98% and accurately deduced that they were made with Suno.
As a cross-check, I also uploaded some old MP3s from my song library, which got a very low AI probability score.
The drawback with IRCAM's tool is that you can't paste links to songs, so you can't check tunes that you can only hear on a streaming service.
There are a few other websites that let you both upload song files and paste Spotify links for analysis, but they have their own limitations. When I tried them out for this story, the results were either inconclusive or flagged some AI songs as human-made and vice versa.
Check the lyrics
AI song tools can churn out both music and lyrics. Many serious users like to write their own words and plug them in because they've discovered that AI-generated lyrics tend to be bad.
Casual users, though, might prefer to just let the machine write them. So bad rhyming schemes or repetitive lyrical structures might be a clue that a song is not man-made. But it's subjective.
Some users report that Suno tends to use certain words in its lyrics like 'neon,' 'shadows' or 'whispers.'
If a song includes these words, it's 'a dead giveaway' that it's AI, said Lukas Rams, a Philadelphia-area resident who has used Suno to create three albums for his AI band Sleeping with Wolves. 'I don't know why, it loves to put neon in everything.'
Weekly
A weekly look at what's happening in Winnipeg's arts and entertainment scene.
No easy answers
AI technology is improving so quickly that there's no foolproof way to determine if content is real or not and experts say you can't just rely on your ear.
'In general, it can be difficult to tell if a track is AI-generated just from listening, and it's only becoming more challenging as the technology gets increasingly advanced,' said Mousallam of Deezer. 'Generative models such as Suno and Udio are constantly changing, meaning that old identifiers – such as vocals having a distinctive reverb – are not necessarily valid anymore.'
___
Is there a tech topic that you think needs explaining? Write to us at onetechtip@ap.org with your suggestions for future editions of One Tech Tip.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Winnipeg Free Press
26 minutes ago
- Winnipeg Free Press
Generations come together at a county fair dubbed Mississippi's ‘giant house party'
PHILADELPHIA, Miss. (AP) — Each summer, hundreds of brightly colored cabins come to life with the sound of children playing and smells of Southern comfort food in what's known as Mississippi's 'giant house party,' the Neshoba County Fair. The fair touts itself as the largest campground fair in the country, where attendees cram into more than 500 two-and-three story wood cabins for eight days every year. The larger cabins can sleep upwards of 30 people, sometimes in the same room. 'It's like having two Christmases a year,' said Mike Hardy, who attends the fair just about every year and shared a cabin this year with 20 members of his family, from infants to grandparents. For Hardy, who lives more than 300 miles (500 kilometers) away in Nashville, Tennessee, the fair is one of the only times he visits his hometown. He calls it a high school, college and family reunion all wrapped into one. 'I wouldn't miss it for anything,' he said. 'It's just always been a big part of our lives.' Hardy inherited the cabin from his father, who bought it in the late 1960s. It's located in what's known as 'watermelon alley,' one of several neighborhoods that divide up the community, which feels like a mix between a candy-colored frontier town and an amusement park. His children grew up going there. The pictures they drew on hot summer days still hang on the walls, joined now by their own children's artwork. His daughter, Madison Hardy-Dennis, attended her first fair when she was less than a year old. Now, her 6-year-old twins run barefoot in the red Mississippi mud, play pranks and get into water balloon fights — just like she did. 'I hope that they understand how special this week is, and that this place is,' Hardy-Dennis said. Horse-race watching at the nearby race track and card playing are among Hardy family's favorite activities during the fair. They take their kids to the carnival rides and cook large family meals. On their way to the track, they walk through Founders Square, the oldest section of cabins with a pavilion used for dances and political speeches. It's where Ronald Reagan gave his famous states' rights speech in 1980 while running for president. Sid Salter, whose family has been going to the fair since it first opened in 1889, said it's a place where children are safe to roam freely. Often, parents write their kid's name and cabin number on their arms. If they get lost, a friendly fair-goer will help them find their way back. The communal atmosphere extends to mealtime. Although only about 20 people stay in their cabin, Salter's family often feeds 50 or 60 people a day. 'It's not an inexpensive hobby,' he joked, 'but it's a great time with people you only see, you know, during the fair.' The fair, Salter said, also feels like a reunion with loved ones who are no longer living. He imagines that the spirits of his twin sister, first wife and parents like to 'knock around' the campground where they made so many memories. 'It may be a figment of a fertile imagination — I'm sure it is — but I feel it,' he said. At 66 years old, Salter has only missed three Neshoba County Fairs, once for an adventure camp when he was 13, again to cover the 2000 Republican National Convention as a reporter and in 2017 when he was battling cancer. He said he often eats the same meals, does the same activities and sees the same people year after year. 'In a sea of change in every facet of our lives, the fair is constant,' he said.
Globe and Mail
4 hours ago
- Globe and Mail
Meta Platforms, Inc. Earnings Call Highlights AI Growth
Meta Platforms, Inc. ((META)) has held its Q2 earnings call. Read on for the main highlights of the call. Elevate Your Investing Strategy: Take advantage of TipRanks Premium at 50% off! Unlock powerful investing tools, advanced data, and expert analyst insights to help you invest with confidence. Meta Platforms, Inc. recently held its earnings call, showcasing a strong performance driven by advancements in AI technology. The company reported significant growth in revenue and user engagement, attributed to its strategic investments in AI infrastructure and talent. Despite these achievements, Meta faces challenges, particularly with losses in its Reality Labs division and regulatory pressures in the European Union. Strong Revenue Growth Meta reported a total revenue of $47.5 billion for the second quarter, marking a 22% increase on both a reported and constant currency basis. This impressive growth highlights the company's robust financial health and its ability to capitalize on market opportunities. AI-Powered Ad Improvements The company has seen a notable enhancement in ad conversions, thanks to AI-powered recommendations. Instagram experienced a 5% increase in ad conversions, while Facebook saw a 3% rise. These improvements underscore Meta's commitment to leveraging AI to optimize its advertising platforms. Increased Engagement Meta's advancements in recommendation systems have led to increased user engagement, with a 5% rise in time spent on Facebook and a 6% increase on Instagram. This growth in engagement is a testament to the effectiveness of Meta's AI-driven strategies. Family of Apps Revenue Surge The Family of Apps division reported a revenue of $47.1 billion, reflecting a 22% year-over-year increase. This surge underscores the strength and popularity of Meta's suite of applications. Strong Operating Income Meta's operating income for the second quarter was $20.4 billion, representing a 43% operating margin. This strong performance indicates efficient cost management and a successful revenue model. Growing Meta AI Engagement Meta AI now boasts over 1 billion monthly active users, highlighting the widespread adoption and integration of AI across Meta's platforms. Expansion of AI Devices Meta is making strides in the AI devices market with the successful launch of Ray-Ban Meta glasses and the new Oakley Meta HSTN. These products are gaining momentum, contributing to the company's growth in the augmented reality space. Reality Labs Operating Loss Despite the overall positive performance, Reality Labs reported expenses of $4.9 billion and an operating loss of $4.5 billion. This division remains a challenging area for Meta, requiring strategic adjustments to achieve profitability. Regulatory Challenges in Europe Meta is facing increased legal and regulatory headwinds in the EU, which could significantly impact its business and financial results. Navigating these challenges will be crucial for Meta's continued success in the region. High Infrastructure Costs The company's total expenses for the second quarter were $27.1 billion, up 12% due to higher infrastructure costs. This increase reflects Meta's ongoing investments in its technological backbone to support future growth. Increase in Capital Expenditures Meta's capital expenditures reached $17 billion, driven by investments in servers, data centers, and network infrastructure. These expenditures are essential for maintaining and expanding Meta's technological capabilities. Forward-Looking Guidance Looking ahead, Meta provided extensive guidance for the second quarter of 2025. The company expects Q3 2025 revenue to range between $47.5 billion and $50.5 billion. Total expenses for 2025 are anticipated to be between $114 billion and $118 billion, driven by continued investments in infrastructure and talent. Meta's strategic focus remains on AI development, with advancements in the Llama 4.1 and 4.2 models and growth in the augmented reality space. In conclusion, Meta Platforms, Inc. has demonstrated strong financial performance and user engagement growth, driven by its strategic focus on AI advancements. While challenges persist, particularly in Reality Labs and regulatory environments, Meta's forward-looking guidance suggests continued growth and investment in key areas. Investors and stakeholders will be keenly watching how Meta navigates these challenges and capitalizes on its opportunities in the coming quarters.
Globe and Mail
5 hours ago
- Globe and Mail
Toronto's Blue J Legal raises $167-million as demand for its ChatGPT-like chatbot for tax pros takes off
Blue J Legal Inc, a Toronto software startup that sells a chatbot for tax specialists powered by large-language models, has secured $167-million in financing led by two U.S. venture capital firms. The company reported in two filings on July 25 with Canadian securities regulators that that it had raised $161-million from investors outside Canada, and $6.4-million more within Canada. A corporate filing by Blue J with the federal government last week indicates the financing values it at more than US$300-million. One of the securities filings reveals the company has added two directors, Cathy Gao – a partner with Sapphire Ventures in San Francisco – and Allen Miller, a partner with Oak HC/FT in New York. A source familiar with the matter confirmed the two U.S. funds co-led the deal. Past investors include Canada's Mistral Ventures, Relay Ventures and Generation Capital, as well as U.S. funds Ten Coves Capital and LDV Partners. The Globe and Mail is not identifying the source as they are not authorized to discuss the matter. Blue J was set to announce its financing next Tuesday and only agreed to share details with The Globe on condition it waited to publish the news three hours after a U.S. publication broke the story. The Globe declined. 'At this time, Blue J cannot confirm any details' until Tuesday, spokeswoman Holland Eichorn said in email Friday. The company is one of several legal software firms bringing AI-driven products to law firms to hasten the automation of often tedious, time-consuming legal work. Many have experienced blistering sales growth, sparking massive investor interest. In June, U.S.-based Harvey AI, a legal AI software firm, raised US$300-million in a deal valuing the company at US$5-billion. That was co-led by venture capital giants Kleiner Perkins and Coatue, four months after raising a similar amount at a US$3-billion valuation. And in June, Burnaby, B.C.-based legal software firm Clio paid US$1-billion for Barcelona-based vLex LLC, maker of an artificial intelligence tool that keep lawyers from accidentally citing fake court rulings. Legal data giant Thomson Reuters Corp., whose controlling shareholder, the Thomson family, owns The Globe and Mail, has also been active in developing and buying AI-powered tools to offer clients. Blue J is a University of Toronto spinout led by Benjamin Alarie, the Osler Chair of Business Law with the school's faculty of law, which began developing artificial intelligence-powered software in the mid-2010s. Its original product helped lawyers and accountants speed-up the process of researching complex taxation matters, powered by predictive analytic software. Blue J raised venture capital and its early product sold decently, counting Canada's big four accounting firms and the Canada Revenue Agency as customers. But 'it was OK, it wasn't awesome, it wasn't a home run,' Mr. Alarie said in an interview early this year. After OpenAI launched ChatGPT in late 2022, the conversational AI chatbot that provides text responses to questions in seconds, Mr. Alarie realized his company could fold generative AI capabilities into its product to answer far more tax research questions. Blue J incorporated OpenAI's large language model technology into its product, relaunching it in mid-2023. That effort produced answers in about 90 seconds, but they were wrong about half of the time, a byproduct of the quirky tendency of LLMs to generate 'hallucinations,' or incorrect information. Blue J set about to improve the product, in part by licensing data from Tax Notes, a leading publisher of tax views and commentary, and benefiting from improvements to successive releases of LLMs. By last December, Blue J could produce answers within 10 seconds that were 95 per cent-plus accurate and enabled users to have technical back and forth conversations about complex tax matters with the program. Some firms found their users could save two hours a week each using BlueJ, which retails for $2,995 each year in Canada per user. 'You just need to be able to articulate a technical tax question clearly and you get a ton of value out of Blue J almost immediately, which is a huge selling feature and why things are going so well for us,' Mr. Alarie said in December. By the end of this year, Mr. Alarie has predicted Blue J will be more than 99 per cent accurate. The speed and accuracy of the product spurred interest as legal and accounting firms – from solo practitioners to large enterprises – flocked to the platform. Blue J was boosted by distribution partnerships with the National Association of Tax Professionals in the U.S. and associations representing chartered professional accountants in Canada and the US. Customers include KPMG, Richter LLP and Baker Tilly. By early this year, Blue J was expanding in the UK market. Annual recurring revenues soared, reaching US$2-million by the end of 2023, and US$9-million a year later from the new product. Sales have continued to expand at a fast clip, clocking in at more than $20-million on an annualized basis now, making Blue J a relatively rare company selling software to enterprises to experience strong sales growth amid an ongoing period of economic uncertainty. 'This is really working,' Mr. Alarie said in January. We have really good product-market fit.' Keeping up with constant tax law changes 'is a never-going-away problem.'
