Artificial intelligence is being used to produce a growing number of songs about the Russia-Ukraine war, with some of the music presenting Russia's invasion in a favourable light and criticising Ukraine.
Get breaking news anytime, anywhere. Download the TOI app now!
Hundreds of such tracks have appeared on major music streaming platforms, including Spotify, Apple Music, YouTube and Deezer, according to research by Terrecon, a Ukrainian nonprofit organisation focused on threat intelligence.
The songs cover a wide range of musical styles, from ballads to heavy metal. Terrecon said some praise Russian soldiers, describe the war from a Russian perspective and contain derogatory references to Ukrainians, including ethnic slurs.
One of the more popular tracks identified by the organisation is “Power of Artillery”, a Russian-language rock song centred on artillery systems including Krasnopol and Carnation.
Nearly 1,000 songs identified
Terrecon said it catalogued 981 songs created with AI that contained what it classified as pro-Russian war propaganda.
Among the songs whose release dates could be established, most appeared after 2024. That coincided with the wider availability of AI music generators such as Sudo and Udio.
The technology has lowered the time and cost involved in producing music, according to Daria Verbytska, a founder of Terrecon.
“They have found new ways to create cheaper drones and use them on a massive scale,” Daria Verbytska, a founder of Terrecon, said told NYT in a telephone interview, referring to the Russians. “It’s like that with music — creating massive propaganda with the help of AI”
Some of the songs use themes associated with Soviet-era war music but present them through contemporary genres. One track, “Soldier's Daughter”, describes the death of a soldier and the impact on his child while also portraying the soldier's sacrifice in a positive light.
It is not clear how much involvement, if any, the Russian government has had in producing or distributing the songs. However, the Kremlin has supported cultural projects aimed at promoting patriotic themes.
Last week, a director of Moscow's Internet Development Institute, an organisation that works closely with the Kremlin to promote patriotic values, said in a presentation that the institute had supported more than 350 artistic projects related to the war, according to Russian media reports.
AI tools can produce war-related music
Terrecon also tested several AI music generators to see how they responded to prompts requesting songs promoting Russian views of the war.
Researchers used Suno, Udio and Google Flow Music in the US, along with Singapore-based Mureka. They entered prompts in both Russian and English, including requests involving violence and derogatory descriptions.
The platforms rejected some prompts, but researchers said they were nevertheless able to generate songs resembling some of those in Terrecon's database. They then uploaded the tracks to YouTube Music and SoundCloud.
“The moderation systems mainly responded to specific banned words rather than the overall meaning of the prompts,” Terrecon’s report found.
The researchers said the issue is partly linked to the differing rules adopted by AI music services and streaming platforms. They also said automated systems may have difficulty identifying expressions, slang and offensive language in different languages.
Streaming platforms tighten AI rules
Major music platforms generally have policies covering hate speech and violent content, although fewer have specific rules dealing with propaganda.
Deezer and YouTube are among the platforms that require certain AI-generated content to be labelled. Spotify said in September that it had introduced a policy to label AI-generated songs and prevent them from being recommended to users. The company also added labels to songs shared by The New York Times as examples.
Terrecon said the songs it identified had been played millions of times across major streaming services, although the extent to which listeners understood that some of the music was AI-generated or politically motivated is unclear.
Deezer spokesman Jesper Wendel said the platform reviews potentially rule-breaking content on a case-by-case basis.
SoundCloud spokeswoman Sade Ayodele said the company reviews reported content and added that “there is no place on SoundCloud for content that incites or glorifies violence.”
The emergence of these songs illustrates how AI music tools can be used to produce and distribute large volumes of war-related material. It also presents streaming and AI companies with the challenge of deciding how existing rules on hate speech, violence, political content and AI-generated material should apply to such music.
