Digital: Disrupted: How is AI Reshaping the Way We Work?
September 22, 2023
In this week’s episode, Paul sits down with Ricardo Michel Reyes for a conversation around how AI is impacting and shaping the employee experience. As an expert in AI, Ricardo shares what he would say to skeptics of the technology, and how HR leaders can implement AI.
Digital: Disrupted is a weekly podcast sponsored by Rocket Software, in which Paul Muller dives into the unique angles of digital transformation — the human side, the industry specifics, the pros and cons, and the unknown future. Paul asks tech/business experts today’s biggest questions, from “how do you go from disrupted to disruptor?” to “how does this matter to humanity?” Subscribe to gain foresight into what’s coming and insight on how to navigate it.
About This Week’s Guest:
Ricardo Michel Reyes is the Co-Founder and Chief Science Officer at Erudit AI, a people-first AI for HR. He is also a member of the IA2030MX coalition, which led to the opportunity to educate policymakers on AI as they developed the National Artificial Intelligence Strategy of Mexico.
Listen to the full episode here or check out some highlights below.
Paul Muller: The thing that was most recently hyped right ahead of AI was machine learning. In your mind, is there a difference between machine learning and AI, or are they just variations on the same theme?
Ricardo Michel Reyes: That's a good question. So, humans are the ones who make the labels for everyone. Nature just is there, right? Science, for example, used to be called natural philosophy and chemistry, biology, geology, everything was natural philosophy and as it progressed, we started dividing and dividing and dividing and dividing, and now you have like 200 branches of physics and 50 branches of chemistry. And so, AI is the same. AI would be the natural philosophy of AI, the very general term that contains any form of learning and symbolic manipulation and structures, that a manipulation and anything that is not, you put X inside, you do some kind of operation, and you take Y outside whenever you are not directly coding what the thing does, but rather learning from the data. So, then machine learning is a big part. Almost every branch of AI uses some form of machine learning, but it's not all of AI. AI contains for example, also computer vision. So, you use also machine learning for computer vision. Both machine learning and computer vision are part of AI. So, you could say that all machine learning is AI, but not all of AI is machine learning.
PM: Talk to me about how you are all thinking about applying AI to the HR and talent management problem.
RR: So, our first approach was to replace service. We hired some really talented psychologists, PhDs in clinical organizational and industrial psychology and had them look at a lot of texts from Slack teams donated by companies or from public data sets to try and see which kind of things they could observe or measure from text. So, then we figured out that we could measure for example, engagement from those texts, that we could measure burnout from those texts, that we could measure if people feel supported by the leaders, if they are coping well with the workload they have. And then we created a product around it to give you all of the metrics that our psychologists analyzed and determined that would be measured from those texts, and models trying to replicate the evaluation of these psychologists. Then we went into the market and realized that no extremes in life are good.
So right now, we're not fighting with service, right now we have a hybrid where we try to provide this software to fill the space between the service because most of the companies have already really talented, really smart psychologists or HR managers from all sorts of backgrounds that already know their people have strategies, already know the problems of the company, already have some solutions in place. But they can only run this service because they're so annoying and because they're so hard to implement, they do it every six months, every year, every two years. So again, AI at scale, what we're doing now is first try and run a survey with the company or to see their latest survey to try to correlate and validate that with the platform and say like, okay, you normally ask these things in your survey.
Look here, these two or three metrics can tell you automatically from Slack, from Teams, from the suite of your employees, the same metrics. So, during the time that you are not doing this service, you can observe the change and the trends in the platform so that your next survey is more optimized for the things that you don't know from the platform. And then we are also learning a lot from all of the things that people ask in their service and from all of the strategies and programs. So, we also keep improving our product to better serve the needs of the customers that we see working with on implementing their programs. This is all based on text. People connect their Slack or their Zoom or their Teams or whatever. And then we send this to networks, and especially transformers that are trained over this psychologist evaluation, expert evaluation. And then we output a score that you can see in the dashboard and you can see the last three days, 60 days, last 90 days to help people have more real time tracking of how their people are feeling and also to be able to see if their strategies are working and to improve the service they normally do every six months or one year.
