Cannes Lions

Fixing the bAIs

MULLENLOWE MENA, Dubai / AURORA50 / 2024

Case Film
Supporting Images
Case Film
Supporting Images
Presentation Image
Supporting Content
Supporting Images
1 of 0 items

Overview

Entries

Credits

OVERVIEW

Background

Situation:

With AI on the rise, a hidden issue emerges: women are overlooked in a future shaped by AI. It tends to reproduce human biases, particularly against women. AI, powered by biased datasets, can perpetuate gender stereotypes and inequalities, affecting women's representation and opportunities, and regressing years of progress for women in the workplace.

Data from MIT and UNESCO state that AI image-generating tools think 97% of professionals are men.

Brief/ Objectives:

Aurora50, a leading Diversity, Equity, and Inclusion (DEI) Impact Agency in GCC, that promotes inclusive workplaces and diverse leadership, aimed to combat bias, not just in real-world workplaces but also in the virtual world by creating a International Women’s Day campaign that ignites passion for our vision of an inclusive future.

Idea

Fixing the bAIs.

Using AI to fix AI itself.

Using image-generating tools like Midjourney, DALL-E2 and Stable Diffusion, We created a huge databank of inclusive and diverse images of women in various professions. We removed gender from their metadata (file names, tags, and descriptions) and added this powerful database to fixingthebias.com, Using an algorithm our website automatically uploaded this huge database to all the open datasets that are used to train AI and to stock image banks, social media, and the datasets used to train AI. With those images, we started teaching AI that all professionals can be women, too. We turned the tables on AI using its own tools and injected diversity directly into AI's veins.

Strategy

In meticulously gathering data from MIT and UNESCO, we uncovered a concerning truth: AI feeds on biased datasets, amplifying existing anti-female biases is why AI image-generating tools perceive 97% of professionals as men.

Our campaign targeted influential figures—key players in the AI industry, policymakers, AI developers, tech leaders, organisations, media, societal representatives - who could spark meaningful change.

Operating across diverse platforms, from social media to open datasets, we sought to rectify biases within AI. Our impactful online video initiated a global conversation, grabbing the attention of both media and key figures in the AI industry.

Our approach involved generating thousands of inclusive images, strategically populating open datasets to reshape AI's perception of women in various professions Notably, we actively participated in the only Plenary meeting of the CAI, thematically linked to examining the draft Council of Europe [Framework] Convention on AI, Human Rights, Democracy, and the Rule of Law.

Execution

Fixing the bAIs. We used AI to fix AI itself.

Using image-generating tools like Midjourney, DALL-E2 and Stable Diffusion, We created a huge databank of inclusive and diverse images of women in various professions. We removed gender from their metadata (file names, tags, and descriptions) and added this powerful database to fixingthebias.com, Using an algorithm our website automatically uploaded this huge database to all the open datasets that are used to train AI and to stock image banks, social media, and the datasets used to train AI. With those images, we started teaching AI that all professionals can be women, too.

We launched the campaign on International Women's Day on social media with the online video promoting the project, it instantly caught the attention of leading figures in the AI industry, data scientists, and gender equality advocates sparking a global discussion and ensuring the campaign's effectiveness and impact. Additionally, collaborations were established with major stock image companies to incorporate the campaign's images into their datasets, reaching a broader audience and enhancing the inclusiveness of AI models globally.

The significant scale of the campaign extended beyond traditional advertising, aiming for widespread impact and systemic change inspiring 52 countries to create new AI laws and policies and contributing to the passing of the EU's AI Act, influencing gender-conscious AI laws on a global scale.

The campaign, aligned closely with Aurora50's brand values of promoting diversity, equity, and inclusion in the workplace. The campaign exemplified the brand's commitment to challenging gender biases and driving real change in societal perceptions.

We started from initial research and data collection to the creating our own database, dissemination of the database, and injecting our datasets into open datasets used to train AI. Along with ongoing engagement with our target audience.

Outcome

Fixing the bAIs didn't just address the problem but created a long and lasting solution.

The campaign contributed to passing of first regulation on Artificial Intelligence, the EU's AI Act, which deems bias in datasets unacceptable, adopted by 52 countries at the EU parliament.

Its Being used as a benchmark to fix gender bias in AI at world summits and councils-Council of Europe and CAI.

Acknowledged by AI scientists at Ceimia and Mila + inspired white papers affirming its significance.

6000% Increase website traffic.

Emerging from a small startup in the Middle East, Aurora50 now stands as a global beacon of change, igniting one of the most significant dialogues on gender equality in the workplace with the highest echelons of political power, through the innovative use of AI.

Aurora50 is not only revolutionising workplaces but also reshaping global perceptions and policies, marking a significant stride towards inclusivity and equality worldwide.

Similar Campaigns

1 items

2 Dubai Lynx Awards
Fixing the bAIs

MULLENLOWE MENA, Dubai

Fixing the bAIs

2024, AURORA50

(opens in a new tab)