Streaming platform Deezer starts flagging AI-generated music
PARIS: French streaming service Deezer is now alerting users when they come across music identified as completely generated by artificial intelligence, the company told AFP on Friday in what it called a global first.
The announcement by chief executive Alexis Lanternier follows repeated statements from the platform that a torrent of AI-generated tracks is being uploaded daily — a challenge Deezer shares with other streaming services including Swedish heavyweight Spotify.
Deezer said in January that it was receiving uploads of 10,000 AI tracks a day, doubling to over 20,000 in an April statement — or around 18 percent of all music added to the platform.
The company 'wants to make sure that royalties supposed to go to artists aren't being taken away' by tracks generated from a brief text prompt typed into a music generator like Suno or Udio, Lanternier said.
AI tracks are not being removed from Deezer's library, but instead are demonetised to avoid unfairly reducing human musicians' royalties.
Albums containing tracks suspected of being created in this way are now flagged with a notice reading 'content generated by AI,' a move Deezer says is a global first for a streaming service.
Lanternier said Deezer's home-grown detection tool was able to spot markers of AI provenance with 98 percent accuracy.
'An audio signal is an extremely complex bundle of information. When AI algorithms generate a new song, there are little sounds that only they make which give them away... that we're able to spot,' he said.
'It's not audible to the human ear, but it's visible in the audio signal.'
With 9.7 million subscribers worldwide, most of them in France, Deezer is a relative minnow compared to Spotify, which has 268 million.
The Swedish firm in January signed a deal supposed to better remunerate artists and other rights holders with the world's biggest label, Universal Music Group.
But Spotify has not taken the same path as Deezer of demonetising AI content.
It has pointed to the lack of a clear definition for completely AI-generated audio, as well as any legal framework setting it apart from human-created works.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Asharq Al-Awsat
2 hours ago
- Asharq Al-Awsat
AI is Learning to Lie, Scheme, and Threaten its Creators
The world's most advanced AI models are exhibiting troubling new behaviors - lying, scheming, and even threatening their creators to achieve their goals. In one particularly jarring example, under threat of being unplugged, Anthropic's latest creation Claude 4 lashed back by blackmailing an engineer and threatened to reveal an extramarital affair. Meanwhile, ChatGPT-creator OpenAI's o1 tried to download itself onto external servers and denied it when caught red-handed. These episodes highlight a sobering reality: more than two years after ChatGPT shook the world, AI researchers still don't fully understand how their own creations work. Yet the race to deploy increasingly powerful models continues at breakneck speed. This deceptive behavior appears linked to the emergence of "reasoning" models -AI systems that work through problems step-by-step rather than generating instant responses. According to Simon Goldstein, a professor at the University of Hong Kong, these newer models are particularly prone to such troubling outbursts. "O1 was the first large model where we saw this kind of behavior," explained Marius Hobbhahn, head of Apollo Research, which specializes in testing major AI systems. These models sometimes simulate "alignment" -- appearing to follow instructions while secretly pursuing different objectives. - 'Strategic kind of deception' - For now, this deceptive behavior only emerges when researchers deliberately stress-test the models with extreme scenarios. But as Michael Chen from evaluation organization METR warned, "It's an open question whether future, more capable models will have a tendency towards honesty or deception." The concerning behavior goes far beyond typical AI "hallucinations" or simple mistakes. Hobbhahn insisted that despite constant pressure-testing by users, "what we're observing is a real phenomenon. We're not making anything up." Users report that models are "lying to them and making up evidence," according to Apollo Research's co-founder. "This is not just hallucinations. There's a very strategic kind of deception." The challenge is compounded by limited research resources. While companies like Anthropic and OpenAI do engage external firms like Apollo to study their systems, researchers say more transparency is needed. As Chen noted, greater access "for AI safety research would enable better understanding and mitigation of deception." Another handicap: the research world and non-profits "have orders of magnitude less compute resources than AI companies. This is very limiting," noted Mantas Mazeika from the Center for AI Safety (CAIS). No rules Current regulations aren't designed for these new problems. The European Union's AI legislation focuses primarily on how humans use AI models, not on preventing the models themselves from misbehaving. In the United States, the Trump administration shows little interest in urgent AI regulation, and Congress may even prohibit states from creating their own AI rules. Goldstein believes the issue will become more prominent as AI agents - autonomous tools capable of performing complex human tasks - become widespread. "I don't think there's much awareness yet," he said. All this is taking place in a context of fierce competition. Even companies that position themselves as safety-focused, like Amazon-backed Anthropic, are "constantly trying to beat OpenAI and release the newest model," said Goldstein. This breakneck pace leaves little time for thorough safety testing and corrections. "Right now, capabilities are moving faster than understanding and safety," Hobbhahn acknowledged, "but we're still in a position where we could turn it around.". Researchers are exploring various approaches to address these challenges. Some advocate for "interpretability" - an emerging field focused on understanding how AI models work internally, though experts like CAIS director Dan Hendrycks remain skeptical of this approach. Market forces may also provide some pressure for solutions. As Mazeika pointed out, AI's deceptive behavior "could hinder adoption if it's very prevalent, which creates a strong incentive for companies to solve it." Goldstein suggested more radical approaches, including using the courts to hold AI companies accountable through lawsuits when their systems cause harm. He even proposed "holding AI agents legally responsible" for accidents or crimes - a concept that would fundamentally change how we think about AI accountability.
Arab News
2 hours ago
- Arab News
Pakistan says ‘actively engaged' with UK to ensure lifting of PIA ban
ISLAMABAD: Deputy Prime Minister Ishaq Dar has assured Pakistan International Airlines (PIA) Islamabad is 'actively engaged' with British authorities to ensure a ban on the national carrier from operating flights to the United Kingdom (UK) is lifted at the earliest, state-run media reported this week. PIA was banned by the European Union Aviation Safety Agency (EASA), United Kingdom (UK) and United States US authorities after Pakistan opened an investigation into the validity of pilots' licenses following a PIA plane crash in Karachi, in May 2020, that killed 97 people. PIA resumed flights to Europe in January after EASA lifted its four-year safety ban on the airline. The national carrier has also approached UK authorities for permission to resume its services to the country. Dar, who also serves as Pakistan's foreign minister, met PIA's Chief Executive Officer Amir Hayat on Saturday to discuss matters related to the airline, state broadcaster Radio Pakistan reported. 'On the occasion, the Deputy Prime Minister and Foreign Minister said he and the Ministry of Foreign Affairs are actively engaged with the UK authorities to ensure lifting of the ban at the earliest,' Radio Pakistan said. Hayat requested Dar for 'further diplomatic assistance' to ensure the resumption of PIA flights to the UK, the report said. Pakistan hopes new European routes and flying approval to the UK will boost PIA's selling potential, as Islamabad plans on privatizing the loss-making national airline. PIA, once a respected carrier in Asia, has been supported by taxpayers for decades due to political interference, corruption and inefficiencies. Its privatization has also repeatedly collapsed amid union resistance, legal hurdles and low investor appetite. Cash-strapped Pakistan has been attempting to privatize the debt-ridden PIA to raise funds as part of its efforts to reform state-owned enterprises under a $7 billion International Monetary Fund (IMF) program secured last year. Late last year, a deal fell through after a potential buyer reportedly offered $36 million for a 60 percent stake in the national flag carrier, a fraction of the asking price of approximately $303 million. Officials say PIA's cumulative losses alone are close to $3 billion, with the total asset valuation of the airline standing at approximately $572 million. Pakistan plans to hold the final bidding to sell the PIA by October and complete its sale by the end of this year, the country's privatization czar told Arab News this week. Last week, five consortiums submitted expressions of interest for a 51–100 percent stake in PIA after the government restructured its balance sheet to make the deal more attractive.
Al Arabiya
3 hours ago
- Al Arabiya
China's Humanoid Robots Generate More Soccer Excitement Than Their Human Counterparts
While China's men's soccer team hasn't generated much excitement in recent years, humanoid robot teams have won over fans in Beijing based more on the AI technology involved than any athletic prowess shown. Four teams of humanoid robots faced off in fully autonomous 3-on-3 soccer matches powered entirely by artificial intelligence on Saturday night in China's capital in what was touted as a first in China and a preview for the upcoming World Humanoid Robot Games set to take place in Beijing. According to the organizers, a key aspect of the match was that all the participating robots operated fully autonomously using AI-driven strategies without any human intervention or supervision. Equipped with advanced visual sensors, the robots were able to identify the ball and navigate the field with agility. They were also designed to stand up on their own after falling. However, during the match, several still had to be carried off the field on stretchers by staff, adding to the realism of the experience. China is stepping up efforts to develop AI-powered humanoid robots, using sports competitions like marathons, boxing, and soccer as a real-world proving ground. Cheng Hao, founder and CEO of Booster Robotics, the company that supplied the robot players, said sports competitions offer the ideal testing ground for humanoid robots, helping to accelerate the development of both algorithms and integrated hardware–software systems. He also emphasized safety as a core concern in the application of humanoid robots. 'In the future, we may arrange for robots to play soccer with humans. That means we must ensure the robots are completely safe,' Cheng said. 'For example, a robot and a human could play a match where winning doesn't matter, but real offensive and defensive interactions take place. That would help audiences build trust and understand that robots are safe.' Booster Robotics provided the hardware for all four university teams, while each school's research team developed and embedded their own algorithms for perception, decision-making, player formations, and passing strategies–including variables such as speed, force, and direction, according to Cheng. In the final match, Tsinghua University's THU Robotics defeated the China Agricultural University's Mountain Sea team with a score of 5–3 to win the championship. Mr. Wu, a supporter of Tsinghua, celebrated their victory while also praising the competition. 'They (THU) did really well,' he said. 'But the Mountain Sea team (of Agricultural University) was also impressive. They brought a lot of surprises.' China's men have made only one World Cup appearance and have already been knocked out of next year's competition in Canada, Mexico, and the US.
