The AI community talks a lot about model transparency and accuracy, but I see less discussion around model robustness. Which is a shame. An AI system that changes outputs depending on the day and the phase of the moon is useless to most organizations. Just as we expect our emails to operate accurately and our cars to drive reliably, we expect our AI models to operate reliably and safely -- we expect our systems to be robust. Systems that only function under ideal conditions are not useful for some organizations that need AI models that can scale with and adapt to changes.

As organizations embrace AI, they often handle large volumes of data that power AI systems. Handling data appropriately includes implementing adequate privacy policies and security measures to protect it. Doing so prevents accidental exposure and ensures ethical data use.

The world of health care research provides a great blueprint for AI oversight approached through an institutional review board and a disclosure review board. Learn more.

As companies continue to evolve their recruitment processes and there has been limited research into the perspectives of HR practitioners who actually work with AI-enabled recruiting tools. In fact, studies show that only 32% of companies include individual practitioners in discussions around talent acquisition technology.

This was the focus of my graduate thesis work, which aimed to identify whether the perspectives of HR practitioners on AI-enabled recruiting tools differed from those of HR executives. Based on the findings, here are 7 recommendations for HR executives to ensure their organizations are using AI tools well.

This blog introduces the steps necessary to lay out a practical framework for organizations adopting GAI tools. [repost from SAS Blogs]

I am honored to be recognized as an Innovative Duke Woman on International Women's Day 2023. Click below to see the publication and read about the other women recognized!

Medical practitioners and mental health professional report using ChatGPT to take notes, create treatment plans, diagnose conditions, and communicate with patients. Here are a few questions which you should consider asking your doctors, therapists, or any other health professionals.

Tune into an episode of the Tiflo.info podcast which discusses AI.

In twenty years, you may get an invitation to a wedding where the bride is not human. You will pull out your best cocktail attire, purchase a gift from the registry, and attend a beautiful ceremony. During the reception, you will great the happy couple and wish them the best. The bride will smile politely, joke about an old memory you share, and continue the rounds. You will sit back at your table. It’s just another human-robot wedding.

Interested in the national security implications of artificial intelligence? Enjoy a 12-minute overview that I created as a part of my studies at Duke Initiative for Science & Society and Duke University Sanford School of Public Policy.

Apple Card never intended to be sexist. Nor was it intentional about preventing sexism.

Imagine that you decide to try therapy. Your job is stressful, your family life unsatisfying, and you want to brainstorm with an impartial listener. The convenience of remote therapy is appealing, so you register for a therapy website. Within an hour of registration, you get a message: "Hi—my name is So-and-so. I'm your therapist and I'm here to help you. Let me know when you want to talk!" An hour later So-and-so follows up: "Just checking in. I'm here for you if you want to talk!"

Except they aren’t.