Black founders in the artificial intelligence (AI) space are creating their own versions of AI models to cater specifically to Black and brown communities. OpenAI’s ChatGPT, one of the world’s most powerful AI tools, struggles with cultural nuance and has a bias towards Eurocentric and Western perspectives. John Pasmore launched Latimer.AI to provide answers tailored to reflect the experiences of Black and brown people, while Erin Reddick created ChatBlackGPT, a chatbot focused on Black culture. Spark Plug, based in Canada, offers a ChatGPT for Black and brown students. In Africa, numerous language models have emerged to cater to the continent’s diverse languages and dialects. These efforts aim to ensure that the AI models represent the experiences and knowledge of Black and brown communities accurately.
Personalized AI is crucial because generalized AI models cannot easily capture the African American experience. Many aspects of African American culture are not found online, as they are passed down orally or firsthand. By using sources like the Amsterdam News, a Black newspaper, John Pasmore built Latimer.AI with a focus on accuracy rather than training on user-generated data scraped from the internet. This led to differences between Latimer.AI and ChatGPT in their responses to questions about topics like the Underground Railroad. Erin Reddick’s ChatBlackGPT and Tamar Huggins’ Spark Plug also offer tailored experiences for Black and brown communities.
The development of inclusive AI models is not limited to the African American experience. Africa itself is often overlooked in the AI movement, with a lack of representation in AI journals and a limited number of AI tools supporting the continent’s languages. Yinka Iyinolakan created CDIAL.AI, a chatbot that can speak and understand nearly all African languages and dialects. It focuses on capturing speech patterns rather than text, as many African languages and traditions are communicated orally rather than in written form. CDIAL.AI aims to bridge the knowledge gap by incorporating insights from native speakers and linguists across Africa.
Efforts to create inclusive AI extend beyond chatbots. Steve Jones founded pocstock to address the shortage of minorities represented in stock imaging. pocstock collects diversity data and uses a visual tagging system to contribute to a more inclusive database that businesses can use to train their AI models. While there is still progress to be made, larger stock imaging companies are taking steps to increase the diversity of their content.
The future of AI lies in personalized models that interact with users and understand their individual wants and needs. Black-owned alternatives to mainstream AI models continue to emerge, reflecting the vast and nuanced world we live in. Black founders are hoping to seize opportunities in the AI space, not just as consumers, but as producers and creators, ensuring that AI serves diverse communities accurately and ethically.
