Berliner Boersenzeitung - 'Easiest scam in the world': Musicians sound alarm over AI impersonators

EUR -
AED 4.236015
AFN 75.551379
ALL 93.057443
AMD 422.178264
AOA 1057.706693
ARS 1728.28
AUD 1.635579
AWG 2.076197
AZN 1.959441
BAM 1.953848
BBD 2.325087
BDT 142.637512
BHD 0.435378
BIF 3448.230118
BMD 1.153443
BND 1.477256
BOB 13.737277
BRL 5.892878
BSD 1.154352
BTN 109.953349
BWP 15.551869
BYN 3.431986
BYR 22607.47819
BZD 2.321691
CAD 1.607213
CDF 2618.890393
CHF 0.935736
CLF 0.026821
CLP 1055.596554
CNY 7.782974
CNH 7.782722
COP 3622.560019
CRC 524.092359
CUC 1.153443
CUP 30.566233
CVE 110.15496
CZK 24.250038
DJF 205.564651
DKK 7.475613
DOP 67.304808
DZD 153.213197
EGP 57.981957
ERN 17.301641
ETB 186.667692
FJD 2.551934
FKP 0.852829
GBP 0.854182
GEL 3.010336
GGP 0.852829
GHS 13.552462
GIP 0.852829
GMD 84.779951
GNF 10138.004449
GTQ 8.808048
GYD 241.758495
HKD 9.050506
HNL 30.944027
HRK 7.533153
HTG 150.942487
HUF 365.817254
IDR 20579.725809
ILS 3.464625
IMP 0.852829
INR 110.085029
IQD 1512.289294
IRR 1585666.604091
ISK 142.196729
JEP 0.852829
JMD 183.332423
JOD 0.817808
JPY 183.840894
KES 149.220988
KGS 100.868705
KHR 4680.425919
KMF 491.367
KRW 1634.987876
KWD 0.35638
KYD 0.962039
KZT 538.192371
LAK 26070.956337
LBP 103375.580761
LKR 386.893512
LRD 208.354638
LSL 18.678541
LTL 3.405817
LVL 0.697706
LYD 7.356651
MAD 10.764946
MDL 20.063045
MGA 4944.18958
MKD 61.463601
MMK 2421.58125
MNT 4146.39691
MOP 9.328681
MRU 46.30315
MUR 54.257544
MVR 17.820789
MWK 2001.752407
MXN 19.778773
MYR 4.719904
MZN 73.710788
NAD 18.678541
NGN 1571.530811
NIO 42.478487
NOK 10.931079
NPR 175.936586
NZD 1.964388
OMR 0.4435
PAB 1.154347
PEN 3.899305
PGK 5.10399
PHP 70.667964
PKR 320.487337
PLN 4.300207
PYG 6871.165813
QAR 4.208196
RON 5.241188
RSD 117.358193
RUB 94.582717
RWF 1695.857377
SAR 4.319939
SBD 9.303319
SCR 15.828578
SDG 692.647756
SEK 10.970042
SGD 1.477762
SLE 28.373415
SOS 659.76095
SRD 43.543039
STD 23873.936439
STN 24.475974
SVC 10.10091
SZL 18.675425
THB 38.240663
TJS 10.660381
TMT 4.048584
TND 3.38392
TRY 55.065771
TTD 7.830178
TWD 37.224487
TZS 3056.621
UAH 51.784117
UGX 4299.704797
USD 1.153443
UYU 46.503344
UZS 13774.24018
VES 871.72867
VND 30146.956822
VUV 137.084112
WST 3.153204
XAF 655.319411
XAG 0.017819
XAU 0.000264
XCD 3.117237
XCG 2.080522
XDR 0.815007
XOF 655.305221
XPF 119.331742
YER 274.984033
ZAR 18.746211
ZMK 10382.370012
ZMW 21.592524
ZWL 371.4081
  • CMSC

    -0.1738

    21.57

    -0.81%

  • NGG

    -1.4000

    79.48

    -1.76%

  • CMSD

    -0.1300

    21.69

    -0.6%

  • RYCEF

    -0.2800

    20.57

    -1.36%

  • RBGPF

    -0.6200

    69.88

    -0.89%

  • RIO

    0.8100

    101.91

    +0.79%

  • VOD

    -0.4400

    15.75

    -2.79%

  • BCE

    -0.2100

    22.54

    -0.93%

  • RELX

    0.1000

    35.62

    +0.28%

  • GSK

    -0.8000

    52.16

    -1.53%

  • AZN

    0.4900

    161.91

    +0.3%

  • BCC

    -1.8500

    84.75

    -2.18%

  • JRI

    -0.0800

    12.73

    -0.63%

  • BTI

    -2.2800

    57.05

    -4%

  • BP

    1.2500

    42.88

    +2.92%

'Easiest scam in the world': Musicians sound alarm over AI impersonators
'Easiest scam in the world': Musicians sound alarm over AI impersonators / Photo: Oli SCARFF - AFP

'Easiest scam in the world': Musicians sound alarm over AI impersonators

Does the latest track by your favourite singer sound slightly off? You may be right. Fraudulent AI-generated tracks are increasingly appearing in artists' own profiles on streaming platforms, presented as their original work.

Text size:

British folk musician Emily Portman got a shock in July when she received a message from a fan congratulating her on her new album -- even though she hadn't released one since 2022.

That's when she discovered "Orca" on numerous streaming platforms including Spotify and Apple Music.

The titles of the songs resembled something she might have created herself, but "very quickly I recognised it was AI-produced music", she said.

According to the independent artist, the AI behind "Orca" was "trained" on her previous albums, mimicking her folk-inspired instrumentation and lyrics.

"I just felt really uncomfortable and disturbed that people could be going to my profile ... and then think 'wow, what's this?'," she said.

Portman said people were fooled despite the "pristine perfection" of the vocals and "vacuous lyrics".

The musician couldn't identify the perpetrators of the fraud, but believes she knows how they operate.

Scammers claiming to be artists approach distribution companies, which then upload the music online without any identity checks, she said.

- 'Easiest scam in the world' -

On the other side of the world, Australian musician Paul Bender also discovered from early this year that four "bizarrely bad" AI-generated songs had been added to the profiles of his band, The Sweet Enoughs.

He said the streaming industry hadn't kept pace with security measures such as two-factor authentication now widely used in other sectors.

"You just say: 'Yes that's me' ... and upload a song to whoever's profile," he said.

"It's the easiest scam in the world."

After an Instagram discussion, Bender, who is also the bassist for the Grammy-nominated band Hiatus Kaiyote, received hundreds of messages from artists and music fans.

He compiled a list of numerous suspect albums, particularly in the catalogues of deceased artists, such as the experimental Scottish musician Sophie, who died in 2021.

Around 24,000 people signed a petition Bender launched on change.org, including rapper and singer-songwriter Anderson .Paak and singer Willow Smith, urging platforms to step up security.

- Virtually undetectable -

AI-powered music generators such as Suno and Udio have become increasingly refined.

Almost all listeners are now unable to distinguish AI-generated tracks from the real thing, according to an Ipsos study for the French platform Deezer in November.

This has driven success for bands solely created by AI, such as The Velvet Sundown, which has garnered one million subscribers on Spotify, but also led to a rise in fraudulent activity.

"The reason that music was uploaded under her (Portman's) name was essentially to make sure that they could gain royalties from (it)," said Dougie Brown of the industry representative UK Music.

Revenues on the platforms are generally low, but add up thanks to bots that multiply listening streams tenfold, he said.

Portman and Bender, who have not taken legal action, asked the various platforms to remove the offending tracks -- a process that took between 24 hours and eight weeks.

Some countries and states have legislation to protect artists against imitation, particularly in California.

In others, including the United Kingdom, limited copyright leaves artists vulnerable, said Philip Morris of the Musicians' Union.

He said Portman's case showed how AI-generated music was now so sophisticated it could actually be used "to impersonate the original work of a real artist".

Accused of a lack of transparency, Spotify recently announced measures to make the platform more reliable and transparent.

Like its competitor Apple Music, it says it is working upstream with distributors to better detect fraud.

"Across the music industry, AI is accelerating existing problems like spam, fraud, and deceptive content," it said.

Despite her concerns about potential UK legislation that artists say will damage their interests, and fraudsters making a mockery of the "beauty of the creative process", Portman is working on a new album.

"The album that I'm making, it's costing a lot of money ... but for me it's all about those human connections, creativity and teaming up with other amazing creatives," she said.

(G.Gruner--BBZ)