At first, John Pasmore was excited about ChatGPT.
The serial founder had been in the ai space since at least 2008. He recalled the days when experts declared that it would be decades before the world would see anything resembling a ChatGPT. Fast forward: that day has already arrived.
But there is a problem.
ChatGPT, one of the most powerful ai tools in the world, struggles with cultural nuances. That's pretty upsetting for a black person like Pasmore. Indeed, this oversight has drawn the ire of many Black people who no longer saw themselves adequately represented in the algorithms touted to one day save the world. The current ChatGPT offers overly generalized answers to specific questions that cater to certain communities, as its training seems Eurocentric and Western in its bias. This is not unique: most ai models are not built with people of color in mind. But many black founders are adamant about not being left behind.
Numerous Black-owned versions of ChatGPT have appeared in the past year to specifically cater to Black and Brown communities, as Black founders like Pasmore seek to capitalize on OpenAI's cultural slippage.
“If you ask the general model who some of the most important artists in our culture are, they will give you Leonardo da Vinci and Michelangelo,” Pasmore said of ChatGPT. “It's not going to say anything about India or China, Africa or even African Americans, because it has a bias that focuses on the European trajectory of history.”
Then Pasmore launched ai/” target=”_blank” rel=”noreferrer noopener”>Latimer.ai, a language model for providing responses designed to reflect the experiences of Black people and people of color. Erin Reddick started ChatBlackGPT, a chatbot also focused on black and brown communities. At the global level there is the company based in Canada ai/” target=”_blank” rel=”noreferrer noopener”>Spark plug, which is essentially a ChatGPT for black and brown students. Africa is also experiencing major innovation in this space, with the emergence of language models to serve the more than 2,000 languages and dialects spoken on the continent that Western ai models still overlook.
“We are the keepers of our own stories and experiences,” Tamar Huggins, founder of Spark Plug, told TechCrunch. “We need to create systems and infrastructure that we own and control, to ensure our data remains ours.”
<h2 class="wp-block-heading" id="h-personalized-ai-is-here”>Personalized ai is here
Generalized ai models cannot easily capture the African American experience because many aspects of that culture are offline. Current algorithms search the internet for sources, but many traditions and dialects within African American culture are passed down orally or firsthand, leaving a gap between what an ai model will understand about the community versus the nuances of what actually happens. .
This is one of the reasons why Pasmore tried to use sources like Amsterdam News, one of the oldest black newspapers in the US, while creating Latimer.ai, focusing on accuracy rather than training on data generated by the users extracted from the Internet. By doing this, he began to see differences between his model and ChatGPT's.
He once asked ChatGPT about the Underground Railroad, the passage that enslaved black Americans used to travel to the northern states to escape slavery. ChatGPT's model would mention runaway slaves, while Latimer.ai adjusted the wording, referring to the “enslaved” or “people seeking freedom,” which is more in line with what has become more socially attuned to the talk about the formerly enslaved.
“There are some subtle differences in the language the model uses because of the training data, and the model itself only thinks about black and brown people,” Pasmore said.
Meanwhile, Erin Reddick's ChatBlackGPT is still in beta mode and plans to launch on June 16. Her product works just like it sounds: a chatbot where one can ask questions and receive personalized answers about black culture. “The core of what we are doing is truly community-driven,” she said.
She is in the process of developing the tool, asking users how they want it to look and how they want it to act. It is also partnering with educational institutions like historically black colleges and universities (HBCUs) to work with students to teach and have them help train its algorithm. She said she wants to “create a comprehensive learning opportunity for Black people and people of color to have a safe space to explore ai.”
“The algorithm prioritizes Black information sources so that it can speak to a body of knowledge that is more immediately relatable than the average experience,” he told TechCrunch, adding that, like Pasmore's product, technically anyone can use it.
Tamar Huggins created Spark Plug to also offer a more personalized experience to black and brown communities. Its platform translates educational materials into African American Vernacular English (AAVE), the ethnolect associated with African American communities. That dialect is traditionally transmitted orally and firsthand rather than studied and written like standard English, which means the accuracy of an ai model (or person) that learns it. only internet will falter twitter-account-is-tracking-the-cringiest-misuses-of-black-language” target=”_blank” rel=”noreferrer noopener”>in precision. Capturing AAVE accurately is important, not only so that the GPT responds using it, but also so that students can more easily type prompts that will cause the ai to return the results they need.
“By creating content that resonates with Black students, we ensure they see themselves in education, which is critical for high engagement and academic success,” Huggins said. “When given the opportunity, Big tech will almost always prioritize profits over people. So we created our own lane within the ai space.”
Huggins trained his algorithm on texts like Shakespeare, along with the writings of black authors from the Harlem Renaissance and even his teenage daughter's verbiage to capture the essence of AAVE. Huggins also works with educators, linguists, and cultural experts to review and validate Spark Plug results.
Pasmore is also working to expand his company into schools, especially HBCUs, as more students turn to ChatGPT every day to complete their work.
“This is a better ai companion for a lot of the work that black and brown children have to do,” he said.
Uniting the diaspora
Africa is ignored in the current ai movement. For example, only 0.77% of the world's total ai journals come from sub-Saharan Africa, compared to East Asia and North America with 47.1% and 11.6%, respectively, according to a 2023 study. artificial intelligence Index Report. In terms of population, compared to North America, Africa makes up about 17% of the world's population, compared to only 7% for North America. When it comes time to get information and experts on ai, the chances of sub-Saharan research being used are quite low, which could affect the development of global ai tools.
While Africa is seeing tremendous development in creating more inclusive language models that better serve the Black diaspora, at this time, current ai models, from ChatGPT to Gemini, cannot fully support the 2,000+ languages that They are spoken in Africa.
Yinka Iyinolakan created ai/” target=”_blank” rel=”noreferrer noopener”>CDIAL.ai to address this. CDIAL.ai is a chatbot that can speak and understand almost all African languages and dialects, with special attention to speech patterns rather than text.
Iyinolakan echoed on TechCrunch the same sentiment as many Black Americans: that fundamental ai models are drawn primarily from data from the Internet and most widely spoken languages. Like its African American progeny culture, many African languages and traditions are absent from the Internet, as it is a culture historically communicated orally and not in writing. This means that ai models do not have enough information about African cultures to train themselves, leaving a knowledge gap.
For CDIAL.ai, Iyinolakan brought together more than 1,200 native speakers and linguists from across Africa to gather knowledge and perspectives to build what he calls “the world's first voice-first multilingual large language model.” The company plans to expand over the next 12 months to include even more languages and create a model that supports text, voices and images.
You are not alone here. Google recently awarded Kenya-based Jacaranda Health a $1.4 million grant to develop its machine learning services so it can work in more African languages and Intron Health recently raised several million dollars to expand its clinical recognition of speech for the more than 200 accents spoken throughout Africa. .
“Silicon Valley wants to believe that it is the beginning and the end of artificial intelligence,” Iyinolakan said. “But to 'get' ai, which is what all companies have as their north star, they need to include a third of the world's knowledge.”
Going forward
Embracing ai chatbots isn't the only innovation Black founders are trying to tackle.
Steve Jones founded the company. pocstock create archival images of people of color as, for decades, there has been a dearth of minorities represented in archival images. This is one reason why today's models primarily spit out images of white people when users ask them to generate photos of anything from doctors to pop singers.
“All platforms and tools must be trained on comprehensive, racially inclusive, and culturally accurate data, or else we will (perpetuate) the bias issues currently facing our broader society,” Jones told TechCrunch. To address this, pocstock has spent the last five years collecting diversity data and creating its own visual tagging system that contributes to a database that companies use to help train their ai models so they can produce more inclusive images.
However, some improvements are occurring. Jones said he has noticed that larger stock footage companies that source from ai companies are taking more steps to increase the diversity of their content. Pasmore also sees a brighter future ahead, saying that personalized ai is the future anyway and that the more ai models interact with their users, the more it will understand the wants and needs of a specific person, “which, I think , eliminates many prejudices.” .”
There might even be room for more culture-specific ai models in the future, especially as more Black-owned alternatives continue to emerge. After all, the world is vast and more nuanced – there's no point in trying to fit it into a black box.
“My hope is that more founders of color will get involved in developing their own ai platforms or creating new ai-related jobs as soon as possible in this next economic boom,” Jones said. “ai is going to create trillionaires, and I would love to see people of color take the position of producers and not just consumers.”
