In the world of artificial intelligence (AI), there is a new push to make this technology more inclusive and represent diverse cultures. People like John Pasmore and Erin Reddick have launched projects like Latimer.AI and ChatBlackGPT, which provide AI responses tailored to Black and brown communities. 

New Poll Reveals Black Americans 'More Likely' to Encounter Misinformation Online
New poll says Americans are growing concerned about the online use of personal data. Black people are "disproportionately" encountering misinformation online.
(Photo : Emmanuel Ikwuegbu from Unsplash)

Lack of Cultural Nuance in Mainstream AI Models

Black founders are developing specialized ChatGPT versions to offer more personalized experiences, addressing mainstream AI models' lack of cultural nuance.

John Pasmore, a seasoned AI entrepreneur since 2008, was initially thrilled about ChatGPT. He remembered when experts predicted such technology decades away, but that day has arrived.

However, ChatGPT struggles with cultural nuances, frustrating users like Pasmore. Many Black individuals feel unrepresented by these algorithms, which are often Eurocentric and Western-biased.

The responses from ChatGPT are too generalized for specific questions relevant to certain communities. This issue is not unique to ChatGPT. Most AI models overlook people of color.

Determined not to be left behind, many Black founders have launched their own chatbots and ChatGPT versions in the past year, specifically designed to cater to Black and brown communities. Entrepreneurs like Pasmore are seizing the opportunity to address OpenAI's cultural oversight.

Over the past year, many Black-owned chatbots and specialized ChatGPT versions have emerged to serve Black and brown communities. Black founders are taking advantage of OpenAI's cultural oversight.

Pasmore pointed out that when ChatGPT is asked to name important artists, European figures like Leonardo da Vinci and Michelangelo are typically listed.

It rarely mentions artists from India, China, Africa, or African Americans. This bias stems from the AI's Eurocentric training and focus on European history.

Read Also: Bias AI? Study Finds that AI Algorithms Can Identify Someone's Racial Identity Based on X-rays

Launching Latimer.AI

In response, John Pasmore launched Latimer.AI, a language model providing answers that reflect the experiences of Black and brown people, TechCrunch reported.

Erin Reddick introduced ChatBlackGPT. Another chatbot focused on these communities. Generalized AI models struggle to capture the African American experience because much of this culture is not documented online.

These algorithms typically source information from the internet, but many African American traditions and dialects are shared orally or through firsthand experiences, creating a gap in AI understanding.

To address this, Pasmore utilized sources like Amsterdam News, one of the oldest Black newspapers in the US, while developing Latimer.AI. He prioritized accuracy over relying on user-generated data from the internet. This approach led to noticeable differences between his model and ChatGPT.

Passmore recalled instances where people asked ChatGPT about the Underground Railroad, which was the route enslaved Black Americans used to escape to Northern states.

ChatGPT referred to "runaway slaves," whereas Latimer.AI used terms like "enslaved" or "freedom-seeking people," aligning with more socially conscious language when discussing formerly enslaved individuals.

The language differences stem from the training data, with Latimer.AI specifically designed to consider the experiences of Black and brown people. Pasmore emphasized that the model's focus on culturally relevant terminology reflects a deeper understanding of these communities.

Erin Reddick's ChatBlackGPT, currently in beta, will be launched on Juneteenth National Independence Day on June 19. This chatbot is designed to provide tailored responses about Black culture. 

Reddick is building the tool with community input, asking users for feedback on its design and functionality. She is also collaborating with historically Black colleges and universities to involve students in learning and training the algorithm.

Related Article: Racially Biased AI Poses Significant Risks, Particularly in Facial Recognition Tech

Written by Inno Flores

ⓒ 2024 TECHTIMES.com All rights reserved. Do not reproduce without permission.
Join the Discussion