Chiudere Menu
Notizie

Absolut Vodka seeks to end AI bias

Pernod Ricard’s spirits brand Absolut has teamed up with studio Copy Lab to challenge the AI bias in fashion imagery and ensure algorithm training is more diverse.

(Photos by ABSOLUT on Unsplash)

The news follows reports that AI algorithms are being trained on old, biased data, which is perpetuating and also amplifying existing inequalities and outdated and damaging stereotypes in the fashion industry.

The images created with Copy Lab will feature in the second edition of the world’s first printed AI fashion magazine, Copy Magazine, as well as on  one of world’s largest royalty free stock image websites, Unsplash.

Absolut and Copy Lab processed 10,000 image prompts which highlight a diverse representation of fashion, culture, and identity.  It believes that if the amount of representative content being processed increases, the generative AI algorithms will evolve to better understand the true diversity of modern-day fashion, enabling future users of AI to be able to generate content without as much bias.

Despite progress in the fashion industry, “deep-seated human biases” have permeated AI, the brand said, which  are “resulting in its content reflecting outdated and narrow views of beauty and identity”.

Carl-Axel Wahlström, founder of Copy Lab and expert in AI said that research showed 90% of AI-generated content is in danger of under representation, and was “committed to reshaping the narrative surrounding AI bias in fashion”.

He said: “By leveraging technology responsibly, we can redefine beauty standards and promote an inclusive representation of diversity within the fashion landscape.”

Brand ethos

Absolute’s global marketing VP at Pernod Ricard, Debasree Dasgupta, said that the brand had spent forty years advocating for inclusivity, fluidity and diverse representative.

She said the brand’s ethos was “rooted in the concept of mixing—mixing ideas, identities, people, perspectives, and drinks” and as therefore felt “a deep responsibility to confront bias in AI head-on”.

“By challenging these biases, we aim to contribute to the creation of a more equitable and inclusive technology that positively impacts individuals and communities on a global scale”, she said.

Sembra che tu sia in Asia, vorresti essere reindirizzato all'edizione di Drinks Business Asia?

Sì, portami all'edizione per l'Asia No