Recent conversations and discussions around equality, diversity, and representation have put a massive social spotlight on many of the obstacles our world is facing in regard to discrimination and bias. While there’s still so much progress needing to be made towards fixing the systemic nature of these problems, we should all strive to discover ways within our own communities to better support the underrepresented among us.
When it comes to building greater inclusion inside the tech industry, we can try to look at the issue in two very distinct ways: solving for bias with tech and solving for bias in tech.
Solving for Bias with Tech
Designing products to be more inclusive and representative should seem like a practical course of action for many companies, especially considering the benefits of expanding their market share and helping adopt new consumers. With the recent surge of cultural unrest, we’re seeing many businesses racing to adapt and prove they are meeting the needs of a diverse population.
In a recent post, Band-Aid, a nearly 100-year-old brand and the largest adhesive bandage manufacturer, announced that they would now launch a line of their products to match a wider array of skin tones.
This long-standing racial bias that has existed in many consumer products is also prevalent in technology. Many of the technological advances we see today weren’t built for every skin color. These inherent biases will have to be corrected in order to allow for greater equality and serve the larger community.
Racial Bias Was Built into Photography
For decades, photography and other types of image capture had been processed in such a way that would treat lighter skin as “normal”, leaving other tones to require some special treatment or color correction in order to be viewed in the same way. In fact, this traditional calibration method made use of a tool called a “Shirley card”, a technique likely named after the model portrayed in the original image. Take a look:
So a light-skinned, brown-haired woman was used as the standard for which all other skin tones were compared and adjusted for. Sounds crazy by today’s standards, right? Well, eventually in the 90s, the imaging company, Kodak updated their Shirley cards to include three women with different skin and hair tones, which provided a small degree of improvement in a technician’s ability to calibrate tones appropriately. However, this was still a far cry from accurately capturing the myriad of hues that make up the human palette.
Unfortunately, these slightly more diverse standards are not industry standards, even today. By the time the push to digital photography came, much of the racial bias that had been the norm in photography and film continued to persist even within the new technologies and techniques that were being developed. Advancements in capture technology and the broader acceptance and appreciation of diversity still did not improve many of the issues which continue to exist and affect productions.
Today, filmmakers are still having to color correct to inaccurate tones. This 24-color McBeth chart is an industry-standard color checker used for calibration and has only two colors to reference for skin color (a generic “light” and “dark”), this method leaves a vast array of tones out of the mix.
With these issues in mind, our team at MOD is developing tools that have been specifically designed from the beginning to stop reliance on older tech as well as outdated and incomplete datasets. The advancements in artificial intelligence (AI) and machine learning are in large part what makes this possible. Through innovative technologies and features within our software, we’re circumventing many of the long-standing racial biases that exist with the image capture process.
Advanced Color Correction
When images are processed through the MOD platform, our color correction process uses machine learning to better match the image to the visible light spectrum range, going to true neutral as opposed to color correcting an incorrect idea of “neutral”. This enables us to get real human tones that are significantly more accurate.
More Inclusive Data
Libraries like the Open Computer Vision Library (Open CV) can often still falsely identify a person of certain skin tones. MOD’s datasets include a larger variety of images and people, allowing our machine learning algorithms to recognize a wider variety of skin tones.
MOD processing takes advantage of the latest techniques, similar to what was built specifically for movies like Marvel’s Black Panther, which provides a more inclusive and nuanced style of color grading.
The goal of implementing these tools and features is to further eliminate the bias built directly into capture and processing software. This is especially important in an age where new high-end technologies including facial recognition, computer vision, and AI have been automating bias due to human error (whether conscious or not) through inaccurate capture techniques and poor representation inside datasets.
Solving for Bias in Tech
A large part of the reason why we see bias or discrimination existing inside of tech products is likely due to the fact that the industry, as a whole, has had major problems with representation and diversity in its workforce and leadership. Without people from different perspectives, backgrounds, and life experiences all contributing to technology’s development, how can we trust that we’re legitimately future-proofing our needs in a way that will best serve everyone and not just a single portion of the population? Honestly, we can’t… and without a diverse workforce, we won’t.
Get a load of this: Of the 50,000 employees working at Google (US) back in 2014, 83% were men and 60% were white, while only 2.9% were Latino, and just 1.9% were Black.
Since that 2014 report, Google has spent considerable millions to try and create greater diversity within its personnel. However, from numbers shared as recently as May (as reported by the L.A.Times), Google is showing that around 5.9% of its workers are Latino and only 3.7% are Black. And to further drive home this incredible disparity, keep in mind that Google now has more than twice the number of employees than it did back in 2014. This lack of representation, in one of the largest tech companies in the US, illuminates the rather disproportionate rate at which BIPOC individuals are obtaining employment across the tech industry.
Much of the technology and infrastructure being built today will greatly impact large portions of society, and any lack of intrinsic equality can have long-lasting impacts on areas that put minorities and vulnerable communities at risk. Software and AI must be designed with diverse perspectives so that their use can reduce bias and discrimination especially in crucial service areas like policing, finance, education, healthcare, and human resources management — places that have long been viewed as already having critical systemic problems with prejudice.
There’s vital importance to foster symbiosis between inclusion and technology. Tech can only be its best if we work to increase representation, true access, and equality. Through cross-cultural interactions, we can learn what is needed to create a better future for all. Increasing inclusion is a key way to create a balanced and plentiful ecosystem – a network of interconnecting and interacting parts – that thrives together. This starts within to recognize our own contributions, shortcomings, and real next steps.
“The world as we have created it is a process of our thinking. It cannot be changed without changing our thinking.”
― Albert Einstein
MOD Tech Labs is committed to working with industry organizations and leadership to move the needle toward more representative and inclusive work environments. We’re partnering with industry organizations to help create standards that work to eliminate bias. Through supporting industry-wide initiatives and efforts, we can expand opportunities for overlooked and underserved communities within tech.
One of the individuals we’re excited to work alongside is Christopher Lafayette and his Black Technology Mentorship Program.
Christopher is a well-known expert, speaker, and leader in emergent technology fields dealing with spatial computing, artificial intelligence, XR development, climate tech, and immersive medical solutions. He understands, firsthand, the lack of representation present, and the specific needs to create opportunities for Black talent in tech.
The work he’s doing to open doors for Blacks in technology will create unprecedented access to expertise, training, mentors, and more promoting “STEM from the Start” which will be available from Kinder to Career. On the other side of the equation are mentors that will greatly benefit from the exposure and learning they receive from their mentees.
MOD founders, Alex and Tim Porter are contributing as Allies and Mentors to the program, helping to provide resources, curriculum, and leadership to the endeavor. This program will positively affect many thousands of individuals who are looking to be exposed to technology and learn for their future.
MOD also advocates giving voice to creators from overlooked communities that lack representation in the mainstream. We applaud efforts like the Kaleidoscope Fund which uses community provided grant money and contributions to help back creators from disadvantaged groups by providing funds and support to underrepresented filmmakers and creators.
Through supporting Black creators and developers from directly inside our industries, we’ll be fostering stronger work environments and creating opportunities for years to come.
Build a better future by standing together
As global protests and demonstrations continue to bring a refreshed racial awareness to the forefront of everyone’s minds, we can hope to see many organizations making moves to address these problems head-on. But if this moment in history communicates anything, it’s that these problems are undeniably systemic issues affecting the lives, safety, and happiness of millions of our friends, colleagues, family members, and neighbors. It’s going to take listening to voices from all sides in order to root out the causes of bias and start to result in any true change.
Now, we want to know what you’re up to! How are your personal and professional goals affecting change and helping eliminate bias and discrimination? Are you working with a social cause that should be on our radar? Is there a talented creator from an underrepresented group that you think deserves the spotlight? We want to know about it and learn more, so please get in touch with us at firstname.lastname@example.org.
Additional Reading and References
The Best Algorithms Struggle to Recognize Black Faces Equally