Morgan State announces collaboration on racially inclusive AI project

Listen to this article

Maryland Daily Record Staff

With concerns that artificial intelligence can compound bias, Morgan State recently announced a collaboration to design an AI platform that is reflective and representative of minority groups.

On Jan. 25, Morgan State University’s Center for Equitable AI and Machine Learning Systems (CEAMLS) announced a collaboration with Latimer.ai on the first inclusive Large Language Model (LLM) designed to serve AI users through a racially inclusive LLM experience where the cultural and historical perspectives of Black and Brown people are deeply embedded with the platform. Researchers at CEAMLS will conduct quality assurance and quality control testing of the LLM model.

“The Center is proud to partner with Latimer as we strive towards a future where AI transcends its traditional boundaries to capture the complex mosaic of African American history and culture,” said Kofi Nyarko, D.Eng., professor of Electrical and Computer Engineering with Morgan’s Clarence M. Mitchell, Jr. School of Engineering and director of CEAMLS. “Our collaborative efforts will intensify the pursuit of unbiased AI, ensuring that the digital landscape is as diverse and inclusive as the communities we serve. Through this partnership, we are not just shaping technology but crafting a more inclusive digital narrative for generations to come.”

Latimer is considered the premier online AI resource for accurate historical information and bias-free interaction for Black and Brown audiences. Unlike other LLM’s, it uses licensed content from reputable sources including a recently announced partnership with New York Amsterdam News to build its training platform.

The partnership with Morgan is one Latimer’s founder and CEO is eager to see develop.

“As we build out Latimer, we will enable systems that allow for automation and continue to make the content more reliably free of bias and inaccuracies. Latimer is the first of its kind – a model that looks to ensure the representation of Black and Brown people, now and into the future,” added Pasmore.

Established in 2022, CEAMLS researchers focus on facilitating the development, deployment and verification of socially responsible and equitable artificial intelligence systems. Among the primary areas of interest include the development of formal standards for data preparation, model training, and deployment to promote equitable AI. The Center also delves into the intricate intersection of philosophy and science in which the ethics of AI and the impacts of its widespread application are examined.

According to Nyarko the center aims to accomplish its mission through the formalization of testing protocols for new AI innovations that ultimately promote the detection and mitigation of algorithmic bias.

“Initiatives like the Latimer project speak directly to the type of research and impact CEAMLS seeks in our quest to uproot the inherent bias that can be part of the foundational datasets with AI models in which inclusion and representation are not present in the early stages of development,” Nyarko said.