This summer, Howard University and Google Research developed a new initiative, “Project Elevate Black Voices,” to improve the experience of Black people using automatic speech recognition (ASR) technology, such as Apple’s “Siri” and Google’s “Gemini.”
Because of inherent bias in the development process, many ASR technologies generate incorrect results when speakers of African American English (AAE) vocalize commands. This has caused many Black users to inauthentically change their voice patterns away from their natural accents (a phenomenon known as “code-switching”) to be understood by voice products. As a result, there is a significant lack of natural AAE speech found within speech data.
To address the implicit bias barriers in ASR performance, researchers from Howard and Google collected 600 hours of data from users of different AAE dialects from across the country that can be used by artificial intelligence developers to enhance their products and create equitable technology. Named the Howard African American English Dataset 1.0, this collection will initially be made available only to researchers at historically Black colleges and universities. Additionally, developers at Google will use the dataset to improve their existing ASR products.
“African American English has been at the forefront of United States culture since almost the beginning of the country,” said Gloria Washington, associate professor at Howard University and co-principal investigator of Project Elevate Black Voices. “Voice assistant technology should understand different dialects of all African American English to truly serve not just African Americans, but other persons who speak these unique dialects. It’s about time that we provide the best experience for all users of these technologies.”
“Working with our outstanding partners at Howard University on Project Elevate Black Voices has been a tremendous and personal honor,” said Courtney Heldreth, co-principal investigator at Google Research. “It’s our mission at Google to make technology that’s useful and accessible, and I truly believe that our work here will allow more users to express themselves authentically when using smart devices.”
I knew Howard University was desperate for validation and money. This is beyond the pale with one sided partnership with racist a*s Google and their project to be more inclusive for the so-called AAVE. This AAVE narrative is nothing more than intellectual laziness of the highest order. Maybe racist a*s Google decided to pursue this project unbeknownst to the aspiring to validated Black academics at Howard of the numerous lawsuits against Google and their racist a*s algorithms and hiring. Thereby, this project will give Google some more cover in order to continue about their normative behavioral and research practices.
What I found even more alarming was the high number of Third World country immigrants (mostly women and a few men) from the continent of Africa and the Caribbean who comprised this team. Let’s be honest. These people are not experts on so-called Black American speech patterns by any stretch of the imagination. In fact, Google should call this project is geared towards speech patterns for those hailing from the African and Caribbean countries and should be called ACAVE (African and Caribbean Vernacular English).
In the spirit of transparency Howard and Google, we would greatly appreciate your sharing with the American public the amount it costs for this research project. Further, it’s a sad state of research and academic affairs when you culturally questionable people such as Courtney Heldreth (i.e., tragic mulatto), Pan Pan Jiang, Bob McDonald, and Francios Beaufays who literally enrich themselves at the expense of so-called Black Americans. For those who dissent, I challenge you to say “the racism is in the algorithms and software”. The only for this to change is for Whites, Asians, Latinos, and others is to have an immediate mental cleansing of their racist proclivities.
I challenge racist a*s Google to conduct a similar project specifically focusing on “White Americans Vernacular English (WAVE) starting with Whites from Bangor, Maine, Madison, Wisconsin, Des Moine, Iowa, Bozeman, MT, and Terrebonne, Louisiana in order to “make technology that’s useful and accessible to all users”.
Moving forward, this highly questionable project is no more different than with the Slave Narrative project via the National Archives.
Somebody at racist a*s Google and compliant Howard does Google still have their unofficial “Jim Crow” section at the Googleplex-Mountainview, CA location where those affiliated with HBCUs are situated?