Creating an Afrocentrist future for AI

In her 2018 book Algorithms of Oppression: How Search Engines Reinforce Racism, UCLA professor Safiya Noble eloquently states, "Rather than being a neutral arbiter that sorts content by quality, search engines are easily gamed in ways that reflect discriminatory practices" 

Generative AI tools easily fit into this assessment as well. But what if we can change that? 

My firsthand experience with the inherent biases of AI came when my phone categorized my daughter’s 3C/4A textured afro as a hat in a picture. Upon noticing this, I searched Google for "afro hair," and was provided 844 million results.

And yet the image data from the picture resulted in a completely wrong categorization. In this instance, how did the visual data get categorized incorrectly? In the age of AI, there are more questions than answers, and I can't help but feel like we have a small window to get this right before it’s too late.

So I began exploring the intersection between Afrocentricity and generative AI. 

Afrocentricity is a theory of social change that affirms the importance of African history, culture and values as a foundation for understanding current and future experiences. I’d add that the experience evolves to a place of communal love and liberation. 

Octavia Butler, the mother of afrofuturism, said, "You got to make your worlds. You got to write yourself in." What does "writing yourself in" mean now, in the age of generative AI? And as the technology continues to accelerate, can it be a tool in shaping a reality and future that we have not yet imagined?

I think so. Let's explore.

Generative AI can create text, images, videos and other data models in response to human prompts, making it a valuable marketing tool. AI is projected to grow by 42% and have a business impact of $1.3 trillion by 2032. 

Despite its potential, there have been instances where it has caused harm, such as when Google's generative AI tool inaccurately depicted America's "founding fathers" and the Pope as Black women. Dr. Timnit Gebru, founder of DAIR, warns that standalone data solutions are risky if we don't comprehend the communities we impact.

In 2017, four women of color created an innovative experience called neurospeculative Afrofeminism. They used interdisciplinary product design and virtual reality to envision Black people’s futures, designing products like earrings that record police altercations and clothing that prevents facial recognition. They even recreated the salon experience, wherein scalp care was synonymous with Black women pioneering brain optimization — brilliant, considering the scalp protects the brain. 

The experience uses cultural artifacts within the Black community to give viewers a deeper context and understanding of the Black experience. 

Bringing it into 2024, wouldn’t it be cool if a VR set could data-dump the research captured to an LLM or an AI agent to fuel other datasets?

Last year, artist Curry Hackett used Midjourney to create playscapes with Black hair, combs and rollers and a series with Gullah Geechee women weaving wicker homes out of sweetgrass. 

But for a series on watermelons floating down the Anacostia River, he shared the story of not being able to run the prompt. The tool didn’t understand the southern cultural cues he wanted to bring into his art.

Using the above examples, I want to introduce solutions that can write Black people and other marginalized groups into the future, as Octavia Butler implored.

Remember: We are the data. I challenge the idea that data exists outside our experiences. Humans are the data. We have the opportunity to shape AI and feed algorithms by being more human. As cultural beings, we must generate new data by fostering personal connections and creating opportunities for connective improvisation.

Retain Black talent in tech. This is not just a matter of diversity, but of ensuring the development and proper management of generative AI tools. Increasing representation of Black and other underrepresented tech talent is essential to establishing a diverse talent bench with various experiences and perspectives — which is crucial for informing the evolution of generative AI.

Develop an AI Cypher. For hip hop lovers, the cypher was a unique, unprompted moment when rappers would get together in a circle to create. Before hip hop, it was jazz, the blues and African oral traditions. This type of cultural anthropology signifies improvisation to communal expression and preservation. To change the future of AI, the AI Cypher must be accessible to artists, creators, communities, educators, marketers, musicians, researchers and writers.

Creating this task force is a collective effort. As advertising helps shape culture, we need to look inward first to recruit this task force to guide representation in the way of AI.

Danisha Lomax is the head of client inclusivity at Digitas.