Our social responsibility as Data & AI professionals
Augmenting Capabilities and Empowering Communities Through Knowledge and Technology
Knowledge opens doors, it transforms lives. As Data & AI professionals, we now have the ability to democratize knowledge and data, reaching audiences that previously had no access. By doing so, we can revolutionize their professional landscapes and make a significant impact on their lives, which is our social responsibility.
To delve into the problem-solving process from a cognitive standpoint, let’s break it down. When faced with a problem, we gather real-time data through our senses (i.e., the state of the world). We then draw upon our memory, tapping into past experiences and theoretical knowledge related to the issue. Finally, we synthesize this information and engage in a cognitive process to reason through the combined data and knowledge, arriving at a solution. While this is a simplified overview and often involves iterations, it provides a useful high-level understanding, as we will explore further.
Throughout history, acquiring knowledge has been a challenging and costly endeavor. Education and economic progress are closely intertwined, each influencing the other. In today’s society, information and knowledge are more abundant with each passing generation. Yet, we still witness significant inequalities across different communities. One might expect the internet revolution to diminish these disparities, given that “everything is available on the web.” However, the reality is that the explosion of information has led to widespread ignorance, which appears to be growing. This is a serious issue, as a lack of education impacts every facet of life. If more people had access to more knowledge, they could achieve more, produce more, and even realize their dreams, but unfortunately, this is not the case.
Simply having access to information isn’t sufficient. Knowledge needs to be effectively communicated, which is increasingly challenging as in the modern world attention spans continue to shrink for various reasons. Traditional knowledge mediators like teachers and books (God forbid) are becoming less effective, possibly because teaching methods have remained largely unchanged for centuries.
Let me share a personal story. A few months ago, I had the pleasure of being stuck in traffic for 10 long hours with my son returning home from watching the full eclipse. It was the perfect opportunity for a quiet conversation about his life, and I learned a ton! We discussed his struggles with learning at school. He told me that he can’t grasp anything when the teacher “talks at him,” but he learns effectively when the teacher “talks with him.” In other words, lectures don’t work for him. He loses focus after just five minutes at the most, but a conversation goes a long way. He mentioned that having a private tutor for all subjects or learning in very small groups would be ideal, but he understands that it’s impractical because no school system could ever afford that. “Lectures are a way to broadcast knowledge, and that’s all they can afford”, he told me, but it doesn’t work for me. He needs a personal touch, so to help him survive school, my wife spends countless hours discussing literature and history with him. It works, but it’s not a sustainable long-term solution.
Then I asked, “What if we could use AI to replace mom and have conversations with you about your history assignments? Would you be open to that?” My son was very enthusiastic, so we started an experiment. We created multiple agents and tuned Claude (Anthropic) to speak in a cool, teenage-friendly language, using humor to make the conversations fun. We fed the agents all the school assignments and other relevant material, and my son began using them as tutors. Even this simple agent led to dramatic improvements in my son’s learning experience, and his grades skyrocketed. It proved to me that when communicated effectively, knowledge can be transferred efficiently, even to the most impatient audiences.
At work, we tackle complex problems regularly, using real-time data (state-of-the-world) and the knowledge we’ve acquired about the business or system to reason and find solutions. When we think about it, we rely on very specific knowledge to solve particular problems, but it took us many years to obtain this expertise. This long journey is a barrier that many cannot cross, making what is possible and even simple for us, impossible for them.
What if we could communicate that specific knowledge in the right language, in a short and succinct manner, fusing it with real-time data while executing some basic reasoning to recommend a course of action? Imagine a system that mimics the problem-solving process by combining data (from databases) and knowledge (from a corpus of documentation) to help employees solve problems. Injecting the right knowledge, with the right context to the right people, in a way that resonates with them would enhance their capabilities and enable them to achieve things they never thought possible in the past.
At Walmart Robotics (ASR), we implemented exactly this approach. We developed a Copilot and several recommender systems that enable people to operate complex robotic systems. Without these systems, many of those individuals might never have had the opportunity to work with cutting-edge robotics technology, but now they can. Associates have reported that they are happy, they feel at home. Their engagement with our system has demonstrated their success.
Today, we have the technology to build such augmentation systems that empower people to break glass ceilings. We can open doors and drive transformation. We just need to make the decision. This is our social responsibility.
