Programming Equity: Championing Diversity and Inclusion in Artificial Intelligence

“Come with me if you want to live,”

I said this to my son last month when I returned from AIMed24: I was excited about all the innovations I saw at this conference. He looked at me like I had lost my mind. I immediately felt very old and not cool as I realized my son was unfamiliar with this classic line from The Terminator. This the 1984 blockbuster movie that starred the former Governor of California, Arnold Schwarzenegger (now that’s a career pivot) as a cybernetic assassin sent back from 2029 (which is only 5 years from now) to kill the unwitting Sara Connor, whose unborn son would one day save humanity from extinction by Skynet, a hostile artificial intelligence in a post-apocalyptic future. I won’t spoil the twist, but then the time paradox was my only discomfort with the movie; artificial intelligence was fantasy, and a fear of a time-traveling cybernetic assassin did not keep me up at night.

But that future is now, and Artificial intelligence (AI) is ubiquitous; even if cybernetic assassins are still a fantasy (but do we know what Elon Musk is up to?), we need to be ready, not for cybernetic assassins but for a world of AI. This world holds immense potential for progress and innovation, offering us a hopeful future in healthcare.

I read an article that stated we would be the last generation to know what life was like before social media. My son will be the first to grow up in a world where he regularly talks to an AI companion. Alexa is the constant second opinion in my house, and we must whisper her name lest she answer when unnecessary.

In the Terminator movies, we were given a glimpse of a post-apocalyptic future in which robots had made labor camps of human survivors, enslaving them. This dystopian future, created by James Cameron, was a cautionary tale about the potential dangers of intelligent AI. However, real caution should be placed on not bringing old prejudices, practices, and biases into the new world.

Did it make sense for robots to enslave humans who needed to eat, sleep, and who got sick – in a robot society, that is not logical for humans or robots. But the movie was written by 1984 James Cameron, who was brainwashed by old-world American prejudices, where human slaves were a reality in our past.

Already, we see old-world imbalances, such as racial and gender biases, causing AI bias and leading to disparities in new-world healthcare. These biases are often embedded in the data used to train AI algorithms, reflecting historical inequalities and perpetuating them in the present.

For example, in 2019, a widely used AI tool designed to identify high-risk patients for enrollment in care management programs was found to prioritize white over Black patients. This was because the algorithm was trained on healthcare cost data as a proxy for healthcare needs. However, Black patients who may have experienced barriers to access had lower healthcare costs even though they often had more significant unmet healthcare needs. The result was that fewer Black patients were flagged as high-risk, leading to inequitable access to critical care resources. This disparity in AI healthcare could lead to preventable deaths and worsen health outcomes for marginalized communities.

And it’s not just Black vs White. It is also male vs female. AI tools used for diagnosing heart attacks often focus on “typical symptoms” like chest pain, which is more common in male heart attack patients. These algorithms were trained predominately on male datasets, ignoring that women often present with atypical symptoms like nausea or fatigue.

This is where representation is crucial. Diversity is not just a buzzword; it’s a necessity. You will never find the answer if you don’t ask the question. We already know that the need for more representation among healthcare practitioners hurts patients. When magnified by AI training data, this can become a mortal blow. Everyone’s voice and experience need to be included in shaping the future of AI in healthcare; otherwise, the outcome will be old-world problems in a new world.

But while I fight to ensure diversity in AI medicine, I fear the backlash from those who mistrust progress and seek to slow down a world that they feel is already moving too fast. Distrust of innovation is also ubiquitous—my mother still uses a flip phone, and how many of us have had to teach reluctant elders how to use home computers?

In The Shawshank Redemption (one of my favorite Stephen King novels; yes, I read, not just watch old movies), the character Brooks, paroled after 50 years of imprisonment, is so fearful of the outside world, which is so different from the one he left, that he takes his own life rather than adapt.

Implementing AI in healthcare is not without its challenges, and it’s essential to understand and address these concerns, empathizing with the audience’s potential fears and reservations.

Black women have often been at the forefront of innovation; Hidden Figures made the world aware of the human computers, the Black women Katherine Johnson, Dorothy Vaughan, and Mary Jackson, who were instrumental in getting us to the moon and back.

Artificial Intelligence is the future, and minority representation and inclusion are imperative to saving lives. We cannot be hidden figures in the revolution, so I say again,

“Come with me if you want to live.”

P.S. This edition of WCGR was written before the tragic shooting of the UHS CEO and is in no way related. Any loss of life is a missed opportunity for growth and change and should be mourned.