10.12-13.22-DigitalTech-Themis-AI

ELAHEH AHMADI: Hi, my name is Elaheh Ahmadi, and I'm the CEO and co-founder of Themis AI. Themis AI was founded by a group of MIT professors and alumni with the mission to empower the world, to create, advance, and deploy trustworthy AI. At Themis we're tackling the problem of AI risk in safety-critical settings.

With the rise of ML and AI in industry, there has been growing need for trustworthy AI, especially in high-risk AI use cases, such as autonomous vehicle, security, financial modeling, and health care. In such use cases, the failure of model can have catastrophic consequences and enormous costs. Based on our research and study, 75% of executives still haven't taken any action to prevent AI risk.

On top of that, there has been more than $13 trillion of lost and untapped revenue due to using risky AIs. And to make it worse, 62% of organizations are facing challenge to acquire talent in AI, to even help them to solve this problem. After more than five years of research at MIT and Themis, we have innovative solutions to tackle the challenge in deploying trustworthy AI. Our technology is robust and reliable and helps you identify when the model should and shouldn't be trusted.

It's fair and ethical, which means we ensure consistent performance on all data types or subgroups within your data. And it's explainable, which means we convert your black box AI model to a glass box. In one of our key studies with our customers, we helped them deploy a risk aware complex autonomous system. Their system consisted of thousands of machine learning models doing subtasks to enable the autonomous system.

However, one of their challenges and blocker was how to develop risk measures to be able-- for each single model-- to enable a safe autonomous system. With Themis, we were able to enable all of their existing models to be risk aware. And by using our technology, a unified platform, we were able to cut down 12 months of R&D into one week of development. At Themis, we're bringing you an end-to-end risk aware AI workflow.

We help you identify and mitigate the risk from early on in the process of machine learning development, which is data collection and cleaning, all the way to deployment at the edge. And we design our product for both executive and managers and engineers. We have a dashboard to help executive managers be able to track the performance of the models, see the model cards, et cetera, et cetera. And we also provide low-level access for engineers to be able to directly integrate Themis into your existing workflow.

One of our products, which is now available for free and open source is a risk aware AI diagnostic tool. CAPSA comes on top of existing workflows and converts your model into a risk aware one, which means on top of the prediction that your model was giving you before, and was designed to provide as an output, you can now get measures such as bias, uncertainty, and label nodes to help you have a deployment of machine learning algorithm that is actually safe. Moreover, in all of our design, we made sure to minimize the overhead for the user.

For example, in order to get all of the benefit of CAPSA, which is our first product, you only need to add two extra lines of code. And by that, you can get all of the benefits and now have a risk aware AI model. We at Themis want to work with your company. We have partnership or proof of concept to enable risk aware and trustworthy AI at your organization.

Thank you so much for your time. I look forward to speaking with you more at the launch exhibition. And please let me know if you're interested in joining our waitlist for CAPSA. Again, my name is Elaheh Ahmadi. Thank you so much for your time and attention.

[APPLAUSE]

SPEAKER 1: Thank you.