Asu professor says AI can detect deceptive language in business leaders better than traditional lie detectors
Yo, check this out. We got an ASU professor dropping some knowledge bombs about how AI is changing the game when it comes to detecting deception in CEOs. This professor from Arizona State University’s W. P. Carey School of Business is saying that artificial intelligence can now analyze CEO speech patterns and catch these business leaders in their lies with a whopping 84% accuracy. That’s crazy, man! And get this, they’re saying that AI is even better than the old school lie detectors. Talk about a game changer.
So, this professor, Jonathan Bundy, is all about exposing the truth. He’s saying that when CEOs use deceptive language, it’s not just harmless talk. It actually has a financial impact on investors and can cause organizations to be overvalued. But here’s the kicker, he’s also found that deception only works for CEOs up to a certain point. Once they start using too much deception, people start getting suspicious and doubting their credibility. So, it’s like a double-edged sword, man.
The professor is hoping that his study will raise the bar when it comes to detecting deception and make people less susceptible to it. And honestly, that’s an important goal because we can’t have these CEOs running around lying to us, messing with our investments and our trust. We need to hold them accountable, man.
Now, the professor was asked if there was a specific case or scandal that sparked his research, and he mentioned the Wells Fargo scandal back in 2016. That was a messed up situation, bro. The financial markets and analysts were caught off guard by that scandal and it got the professor thinking, “How the hell did they miss all the signs?” So, he wanted to see if AI could do a better job than those so-called experts in detecting deception. And it turns out, it can.
The professor did some serious research to trace CEO deceptions. He looked into different areas like psychology, sociology, criminology, linguistics, philosophy, and management. And he found that deceivers have certain speech patterns, like using fewer singular pronouns, inconsistent verb tense, and fewer words referring to sensory experiences. But here’s the cool part, using machine learning, they were able to build on this research and create a model that detects deception with a mind-blowing 85% accuracy. That’s some next-level stuff, man.
But here’s the thing, AI ain’t perfect. It can’t detect deception 100% of the time, and it never will. So, we can’t just blindly trust these AI models. They’re just tools to help us ask the right questions and be skeptical. It’s all about building a healthy dose of doubt, man.
And yeah, there are some dangers with AI taking over. Trust is a big issue, bro. We gotta be cautious and not fully rely on these machines. And there’s always the risk of biases and inaccuracies, especially with specific demographics. The model was trained on communication from mostly white and male CEOs, so it might not be as accurate for everyone.
But hey, there’s some good news. The algorithm they used in the study is available on GitHub for free. So, companies can get their hands on it if they wanna give it a shot. And there’s even a commercial version for those who wanna take it to the next level.
Now, what’s next in terms of research on CEO deception? Well, there’s a lot of opportunities, my friend. They wanna dig deeper into the motivations behind deception, find ways to avoid it altogether, and understand the differences between CEOs who deceive and those who don’t. And here’s something interesting they found, female and older CEOs tend to engage in less deception, while higher-performing CEOs tend to use more deception. They gotta figure out why that’s the case.
And here’s a wild thought, man. If AI can detect deception, that means it can also be used to lie and deceive better. It’s like a never-ending cycle, bro. We gotta keep improving our models and staying one step ahead. We can’t let AI fool us, man. The truth is out there, and we gotta find it.
