| Synopsis | PROBLEMAccording to UNESCO and MIT, highlight that AI, fueled by biased datasets, perpetuates anti-female biases. Tools like Midjourney, DALL·E 2, and Stable Diffusion depict 97% of professionals as men.IDEA & EXECUTIONFixing the bAIs. Using AI to fix gender bAIs in AI.We used AI image generation tools such as Midjourney, DALL·E 2 and Stable Diffusion to build an extensive collection of inclusive images featuring women across professions. By removing gender labels from the metadata, we ensured unbiased representation. This powerful database was then shared on the website fixingthebias.com which then disseminated it to all the open dataset used to train AI. Additionally, A Discord community continuously generates unbiased images, reshaping technology's future.IMPACTFixing the bAIs is recognized as a solution to address gender bias in AI at global summits and councils, inspiring 52 countries to create new AI laws and policies. This initiative establishes a comprehensive global framework for inclusive AI regulations. Notably, the EU's AI Act (world's first AI law), announced on June 16, 2023, acknowledges bias in datasets as an unacceptable risk. Nine datasets, including Liaon 5B, have undergone training, while major stock image companies contribute by incorporating our images for free download. With a reach of 4.4 billion and $4.5 million in earned media, our initiative has gained substantial impact and recognition. |