When I started my career in AI at DARPA 20 years ago, I was passionate about the potential of AI to transform our world. The reality was much less inspiring. Working with models back then meant becoming very familiar with all the ways they fail.
Fast forward to present day and the success of AI is extraordinary to see. AI is powering amazing applications across every industry. But what too often gets overlooked in the exuberant race to “AI all the things” is that the same failure modes still impact models every day.
In 2016, I started Capital One’s AI team, the Center for Machine Learning (C4ML). We developed AI systems for a variety of applications and were very aware of the risks that come with deploying models into high-leverage business areas that affect millions of people. With that in mind, I was one of the first to call on industry to recognize the risks AI creates and to seek solutions so that AI could be deployed in a trusted, responsible manner. The Wall Street Journal, MIT Technology Review, and others covered this call to action (Capital One pursues explainability to guard against bias, The Financial World Wants to Open AI’s Black Boxes).
The natural follow on conversation was always, “ok, so how do we put in place the appropriate guardrails?” Sadly, there were no good answers — it was an unaddressed part of the AI stack. Organizations were stuck trying to cobble together piecemeal solutions for this critical need and siphoning resources away from core business models in the process.
That is why we are taking the next step in enabling people to benefit from AI — today, we formally announce the Arthur platform for Trusted AI. Built by a world-class team of AI veterans, with input from our development partners, and backed by some of the most forward-looking investors in tech, Arthur is the key piece of AI infrastructure that has been missing.
None of this would be possible without our diverse and dedicated group of people working together. Our founding team includes: Liz O’Sullivan, a noted responsible AI advocate with significant industry experience leading commercial adoption of computer vision and NLP; Dr. John Dickerson, a machine learning faculty member at the University of Maryland who completed his PhD studies at Carnegie Mellon and whose work has been funded by Google, Facebook, and DARPA; and Priscilla Alexander who worked with me at Capital One leading the development and deployment of their highest leverage AI projects.
We are fortunate that the need for Trusted AI is recognized by some of the top investors in the world. We are backed by the incredible teams at Index, Work-Bench, Homebrew Ventures and a wonderful group of angels. We could not be more excited to partner with them as we do our part to help positively shape the future of AI.
Details about our new platform are available on our website or feel welcome to contact us at firstname.lastname@example.org.