Artificial intelligence is beginning to define how we interact with one another and the companies we do business with. It’s everywhere, from chatbots to self-driving cars to recommendation engines. And every day it seems like we’re finding new ways to use it—all in the service of making our lives easier.
But for these AI systems to work at the most fundamental level, they need to build trust. Trust is, after all, the foundation of every solid relationship. Without it, businesses aren’t likely to last very long. That’s why we have principles that guide how we design our products. They’re like a moral compass. They hold us to a higher standard. More than that, they empower us to build a safer and fairer world for the end user.
Of course, having principles is one thing. Applying them every single day is another. That’s why we’ve tied them to our corporate values. In this way, they’ve become central to who we are as a company. So, what are these principles, you might be asking yourself? Simple. We’re committed to building AI products that are safe, responsible, and fair. As mentioned, these correspond to our corporate values, specifically: Love People, Build Trust, and Focus on Impact.
Let’s take a look at each of these principles, define exactly what they are, and how they impact our products in real life.
Keeping customers and end users safe should be a function of every AI system. If they aren’t, everything else is irrelevant. Data security and customer privacy are foundational to what we do. That’s why we’ve implemented Privacy by Design standards. We’re also transparent about how we handle data. We never accept, use, or store any personally identifiable information (PII). We do not compromise on our principle of safety, even in instances where a risky approach might yield better results.
We are responsible for the products we design and their outputs. For a user of our products to make responsible decisions within their organizations, we need to understand how and why a prediction was made. This is one reason why it’s critical that our models are explainable. In other words, that we can understand how and why a prediction was made. So, transparency is essential. We build products that complement and enhance your decision-making abilities. Not products that replace them. At the end of the day, because we take responsibility for the things we build, we’ll never create tools that can be misused or abused in any way.
Bias varies across cultures and contexts. This is something we readily acknowledge and attempt to mitigate in the datasets we use and the models we build. We don’t knowingly introduce bias, appeal to relativistic definitions of fairness, or allow fairness to be used as a tradeoff for revenue. We believe our systems should reflect the best that humans can be, and not the preexisting biases we have already fallen prey to.
Hopefully, you understand how our ethical principles inform our products. If you have any questions or would like to learn more, feel free to get in touch. It’s no secret: AI represents an unprecedented opportunity for businesses everywhere to forge stronger, longer-lasting relationships with their customers. Unfortunately, without trust, none of this is possible. With a strong moral framework to operate from, any business can set itself on the right path for its customers.