Music created by generative AI (gen AI) currently demonstrates a 20% accuracy rate when composing music for specific advertising projects.
This is according to a study by the sound brand agency Stephen Arnold Music (SAM) and the sound testing company SoundOut. The research found that humans can still compose more emotionally accurate and engaging music than the AI generation. However, in its current state, AI can be used effectively to assist in the ideation phase.
The researchers gave Stable Audio’s Gen-AI platform four tasks: to produce music that is sentimental/compassionate, inspirational, funny/quirky, and bold/bold. The platform then generated five iterations of each brief that the researchers examined.
Results vary widely depending on the brief. For example, in four out of five iterations of a “bold” brief, the AI was able to produce results deemed commercially acceptable by the researchers. However, he was only able to reach this threshold in one of the 15 other iterations he created in the other three briefs.
In other words, 80% of the time the music doesn’t meet their benchmark for what would be acceptable for commercial use.
Where will AI have the biggest impact on audio advertising?
Briefs work best
Overall, the researchers described that the AI performed “relatively well” and that its targeted performance was “largely successful for most compositions.”
In particular, the AI was able to succeed the most when it was given “consistent” and “well-aligned” briefs and when asked for music that attempted to evoke correlated emotional attributes. When presented with more complex or nuanced briefs, AI is more likely to fail.
“While humans still outperform AI on the emotional front, this study revealed that ‘composing by numbers’ AI is no longer far behind,” said SoundOut CEO David Courtier-Dutton. “The AI wasn’t bad; it just wasn’t that good [as humans]. With some emotional fine-tuning, we expect that AI will at some point in the not-too-distant future match most human composers.
Courtier-Dutton added that AI does not need to understand emotions; instead, he just needs to know how to summon them to people.
“AI can compose music that moves us emotionally,” he continued. “Now it just needs a little more technical empathy to be able to do that with enough precision for commercial use.”
AI uses in audio
AI is already being used in the audio space in many ways. Spotify, for example, is launching “AI DJs”. AI can also help in advanced targeting efforts, hyper-personalized creative, synthetic voice development and many other cases.
As Jason Brownlee, founder of Colourtext, said The media leader last week: “If smart audio creators can find a way to bottle their AI knowledge and scale it into a super efficient, self-learning and self-serving ad production platform, the sky really is the limit.”
SAM and SoundOut currently recommend using AI in the ideation phase when creating sound branding. Chad Cook, SAM’s president of creative and marketing, added: “When we develop commercial-ready music for leading brands […] there are additional considerations for evoking the right emotion at the right time. Performance, emotional timing, production quality, mixing and mastering are all elements where the human touch has a clear impact.
“Combining the capabilities of humans and AI has real potential for sonic branding in terms of efficiency and quality.”
The use of AI in audio will be one of the key topics at The Future of Audio and Entertainment on April 18.
