Suggestions
Benjamin Roome, PhD
Artificial Intelligence Ethicist and Ed Tech Founder
As a founder of Ethical Resolve I focus on the ethics of artificial intelligence and machine learning. We help companies and other organizations to operationalize their values in data ethics.
In my ed tech career I work mostly in the area of digital credentialing. The Badge List Knova Platform is designed to help organizations issue evidence-based digital credentials at scale.
My Philosophy PhD has allowed me to work in the areas I care most about: ethics, epistemology and education. Twitter: @benjaroome Company Websites: ethicalresolve.com badgelist.com
Highlights
I was just reading Workday's new report "Closing the AI trust gap." This jumped out at me:
💥 Data on AI governance and use is not readily visible to employees.💥
👉 Less than half of employees polled have been provided with regular training for AI use and decision making.
👉 Less than a quarter of employees polled have been provided with guidelines on the responsible use of AI.
🌟 GenAI has ushered in an era in which every company is becoming an AI company, although many companies don't see it that way... yet.
🔴 If these companies don't get their proverbial ducks in a row we're going to start to see a lot of negative impacts emerging.
I'll link to the full report in comments. What's your take on this?
We are seeing increasingly rapid deployment of AI systems in the healthcare industry.
A recent piece by Aashima Gupta suggests that better AI systems designed by Google Cloud and others can lead to “Less burnout, more satisfaction” for the world’s 28 million nurses. When it comes to easing the burdens of paperwork, there are few disagreements.
Real healthcare work is not paperwork though…
💥Realizing the promise of AI in healthcare requires another critical component.
In their research, Madeleine Clare Elish, PhD and Elizabeth Anne Watkins, PhD have called this component “repair work.”
👉 Repair work is the human work carried out to ensure that healthcare practices disrupted by AI are effectively repaired so that the benefit of AI systems can be realized.
👉 As we know, AI systems aren’t automatically valuable in and of themselves. Many doctors and nurses are finding problems with them.
👉 Recent articles in the New York Times, Washington Post and Wall Street Journal voice concerns and hesitations on the part of nurses and doctors. (refs in comments)
👏Repair work is necessary to ensure that AI systems deliver on their promise.
We need to see more effort on the part of AI companies to help nurses and doctors carry out the repair work needed for these tools to realize their potential.
As Elish and Watkins note, “Recognizing repair work shifts our focus from those who initiate a project to those whose work and skill is required to make the project work out in the world.”
AI systems cannot function effectively unless they are properly integrated with the processes of human deployment and oversight.
🌟 Good product design is only the beginning. 🌟
The standard AI industry response, “human-in-the-loop,” cannot fully address the problem. The question we need to be asking is:
💡💡💡“How do we build healthy processes that maximize the positive impacts of AI systems while making the lives of healthcare workers and patients better?” 💡💡💡
