Ep. 23: Lorenzo Patelli - AI and Its Ethical Considerations

October 28, 2019 | 20 Minutes

Dr. Lorenzo Patelli, Associate Professor in the School of Accountancy and Interim Director of the Institute for Enterprise Ethics in the Daniels College of Business at the University of Denver, joins Count Me In to talk about the pervasive phenomenon of artificial intelligence (AI) in accounting and finance today. Dr. Patelli has received numerous honors for his teaching excellence and has published several research articles on topics dealing with business ethics. In this episode, he explains how AI, a foundational field of computer science, has an impact on the accounting field because of its contributions to ultimate judgements across an organization. The intersection of process automation and strategic decisions is where management accountants come in as they consider AI and its ethical implications. Listen now to hear more about the accountants role with artificial intelligence.

Contact Lorenzo Patelli:
LinkedIn - https://www.linkedin.com/in/lorenzo-patelli-2abb518/
University of Denver bio - https://daniels.du.edu/directory/lorenzo-patelli/

Daniels College of Business: https://daniels.du.edu/
Institute for Enterprise Ethics: https://daniels.du.edu/iee/


Adam: (00:00)
Hey everybody. Thank you for listening to another episode of Count Me In. I'm Adam Larson sitting alongside Mitch Roshong and our expert guests for this episode is Lorenzo Patelli. Dr. Patelli, is an associate professor in the school of accountancy and interim director of the Institute for Enterprise Ethics in the Daniels College of Business at the University of Denver. He joined us on count me in to talk about the fascinating intersection of artificial intelligence and ethics. Mitch, can you tell us a little more about Dr. Patelli And his expertise in AI and ethics? 
Mitch: (00:37)
Sure, Adam. Lorenzo Patelli also serves on the editorial board of advances in management accounting and has received a number of awards and honors for his teaching excellence in research. He was recently a part of the elevate the ethics of artificial intelligence event at the University of Denver and regularly writes on the topic. Dr Patelli gives listeners a great overview of AI and tells us all why management accountants have the skills necessary to effectively manage the ethical implications of artificial intelligence. Let's go to the discussion. 
Mitch: (01:08)
Artificial intelligence is certainly becoming more pervasive in accounting and finance today. Can you please give us an overview of what exactly artificial intelligence is and what it can do for us? 
Lorenzo: (01:26)
Sure. Artificial intelligence is a field of computer science, particularly this field is concerned with empowering machines to think, behave and act like a han beings. In other words to make sure that machines have intelligence. Yeah. So several different technologies are developed to replicate the han senses and the ability to draw a conclusion or make judgments based on on the census. So for example, humans are able to talk and artificial intelligence develop develop speech recognition tools. Humans are able to move in artificial intelligence, develops motion planning tools through robotics, for example. Humans are able to see, and artificial intelligence develops computer regions. Humans are able to learn from experience. So artificial intelligence develops machine learning. So how does artificial intelligence and how does as computer science develop this technologies? They use different techniques, such as reinforcement learning. So machines learns from executing tasks and make mistakes and so for example, the LinkedIn notifications that we get from like a website like LinkedIn, a social network lately are based on reinforcement learning. So if we pay attention to this notifications or not it's something captured by the machine and it triggers a learning process. Another technique is deep learning and deep learning has to do with artificial neural networks that are meant to replicate the human brain. So there is one input and one output with multiple hidden layers in between in some advance search engines, are based on this technique. Then we have machine learning and the ability to learn from data without being programmed. So there is a training process where we give data to the machine and the machine is trained and that machine is capable of predict. So Google, Gmail or the automated responses that we, use, you know, or texting for example on our smart phones those are based on versions of machine learning. As I said, then we have computer visions and computer vision tools enable autonomous vehicles for example, to capture and understand the surrounding or facial recognition on our game and our phones and finally we have a natural language processes and with this machine so capable to understand and react to human language. So Alexa or Siri are technologies that we use base on NLP. So this, these all these techniques basically in this this methods are used by artificial intelligence to empower machine to examine a phenomenon and initiate a response based on the analysis of if not. 
Mitch: (05:09)
And with all of these capabilities, can you explain how AI has already started playing a role in accounting and finance? 
Lorenzo: (05:15)
Yes. well, specifically management accounting is the practice of judging organizational performance judging, meaning measuring, reporting and interpreting factors that indicate whether an organization is creating good destroying value. So at least in theory, artificial intelligence has and will have their remarkable impact on our profession. because exactly, we deal with analyzing a phenomenon and producing your response, judging it, but also practically, we know that in management accounting is practice relying on machines, databases and these machines nowadays deal with the large quantity of data in complex environments in which processes and products themselves will more and more run on artificial intelligence. So my view is that artificial intelligence will have a huge impact on accounting and finance. we know that companies that are already using some artificial intelligence techniques like bots and machine learning to enter data we know that bots are able to enter and categorize data you fully automated way. We know that something like reinforced learning is assisting companies in internal audit tasks to detect fraud and compliance issues. We know that companies are using language processing techniques to interpret contracts and we can even speculate further and imagine a performance measurement system completed a design based on metrics defined by machine learning techniques. We can again imagine initial iteration of budgets prepare through deep learning and we can envision performance reports and feedback processes, obtain through, language processing techniques. So the impact is going to be a significant and it's going to be a widespread in term of the type of companies who, which are going to be affected and the areas within the companies that are going to be affected. 
Mitch: (07:45)
Now I know another major component of accounting is ethics. So I'm just curious, in your opinion, how do we separate the opportunities from the threats of artificial intelligence when it comes to the ethical application and potentially the negative effects of artificial intelligence and machine learning? 
Lorenzo: (08:05)
Yeah, that's, that's a very, very important question and a strategic finance ad published few interesting pieces on this topic. As far as I see it it's important in first of all to have a framework to look at this issue. And I grew up the unintended consequences of AI in five major buckets. The first one is the impact on work we do know that professions are going to be disrupted and jobs would be lost and economies are actually debating whether artificial intelligence, different from technologists in the past are going to create jobs or destroy jobs primarily. So it's an interesting issue from an economic standpoint and it definitely poses some ethical considerations. The second bucket is inequality. I'm not the official intelligence as specific characteristic relative to a technological innovations of the past that it seems that it's going to benefit primarily some, not everyone. Like I talked earlier about the people that computers electricity to artificial intelligence. Well, electricity was beneficial for the overall population. And so we all benefited from the advancement of this technology. There are some concern, again, it's too early to conclude that artificial intelligence could be controlled and use for the benefits only of a few sections of the population. So that is another thing that we need to be careful about. It's the inequality that could be created by AI. The third bucket is the value transparency. We know very well that you use a jar or two. Digital intelligence are great decisions carry a great deal of biases and discriminatory elements. We all have read several instances on news about imperfect systems that simply replicated human biases, so that is another other area of concern. David Brooks you know, tutorial on the New York times warns about the ethical consequences of artificial intelligence. On the one hand he describes how we have important improvements in observing and preventing and dealing with very delicate problems like mental health issues such as depression and suicides. However David Brooks himself warns the non-transparent use of this data by states, for example, by government and employers. So value transparency is another area where we could see ethical dilemmas caused by artificial intelligence. The fourth bucket is data security and system integrity. so this is very close. It's very connected to management accounting because it has to do with data and there is a growing concern about the safety and security of running and storing data on technologies like blockchain and also the information sharing. So that's the fourth area that I identify a new area of ethical implication of AI. And lastly, I think community and social interaction. So artificial intelligence is going to have an effect on how we relate to each other and for example, the advanced of smart cities full of sensors capable of watching our behaviors constantly. that has also a social implication in which we can find ethical issues. However, I do believe that management accountants should not be scared by the challenge of distinguishing between opportunities and threats because it is our true competency to analyze risk and to measure risk. and so we have again, a great opportunity of contributing to the advancement of this technology, which is gonna affect various aspects of our lives. 
MItch: (12:48)
So as you referenced earlier, artificial intelligence, these bots, they begin to get smarter. Who is ultimately responsible for the decisions that are made by artificial intelligence? 
Lorenzo: (13:02)
Yeah. On this I have a more straightforward answer I guess, which is that the responsibility is also the human being. We cannot attribute responsibility to the machine or a delegate. We should say it responsibility to the machine. So the key is to build people meaning to hire, train and promote people that are competent and principle driven. I have always loved that the competency is the first ethical standard in the IMA statement of ethical professional practice. So management accountants should not withdraw from early discussion and strategic plans and the operational design processes that are ongoing within companies to develop and deploy AI technology. They should use this competency. They should use their competency and contribute to this conversation. So going back to artificial intelligence now if in the past human being succeeded by being a great performance and mostly beating the machine by performing tasks, artificial intelligence I think requires hans to be great persons, meaning we need to rediscover and re-emphasize what makes us unique. And I think what makes us unique as han being is the sense of values. The ability to separate the good and bad, the right and the wrong, the beautiful and the ugly. The Just and the unjust. These is what makes us uniquely human. therefore, to wrap up the answer to your question two pillars. One, the competency that we need to use and contribute with, and the new competency framework for example, of the IMA goes into this direction when it emphasizes technology and analytical skills. and the second pillar, which is strong foundation in professional ethics and values. These two pillars are the key, in my opinion, to enhance the responsibility, which remains, in the hands and of human beings. 
Mitch: (15:20)
So now I'm curious, I'm going to kind of combine two different questions here, but with artificial intelligence, do you believe then there should be some limitations in the capacity that they are used and whether it's yes or no. Do you believe there should be some kind of governing body or oversight body that will set standards for who can use artificial intelligence and how it is used? 
Lorenzo: (15:46)
Yes, this is a very important question and unfortunately the answer is not an easy one and I suspect many attempts will be made and we will inevitably make mistakes and failure and we will fail and we will implement some sort of corrective actions. I am director of the Institute for Enterprise Ethics at the University of Denver and we recently hosted a panel discussion on the ethical implications of AI. I was very surprised by the fact that two two of the panelists were CEOs of startups that are completely based on the development of artificial intelligence two different technologies, two different services. But two companies completely dependent on artificial intelligence. And I was there surprised it said that they both advocated for more regulation on the usage of AI. And I asked myself why, you know, CEOs of startup are so described oftentimes as disruptors and rule adverse you know, people that think beyond the preconceived models sit down on necessarily like regulation. The truth is that in the conversation, what became apparent is that they need framework to operate. They expose themselves to a lot of race and by using this technologies and they welcome a more direction and more regulation, however, I think that the truth as exactly, economists argue it's not so much in preventing all kinds of risks, but what I hold dear is to prevent inequality, more inequality generated by this artificial intelligence. So I think on this, the government could definitely play a role to make sure that we all benefits from these technological innovation. So not just regulation for the sake of regulation, but maybe regulations, my regulation that allows more transparency and more information sharing that would benefit the advancement of this technology for all sections of the population. So it's important to discuss the purpose of this regulation. And then I think the board of directors have a incredibly important role in making sure that organizations are concerned about the ethical implications of AI and they put in place systems and, and how foster cultures we, the in private organization or, or public organization where there is a sensitivity to these issues. So we do have information about positive trends, in this sense. The wall street journal just reported that a venture capital incubators have initiated important changes within start ups is important. Changes are promoting code of ethics that guide artificial intelligence, startups, operations tools to explain how a teacher of intelligence make decisions. And lastly, best practices to communicate and gather feedback about artificial intelligence output. So we see that there is a growing awareness of the importance of getting the board involved in this conversation. 
Announcer: (19:25)
This has been Count Me In, IMA's podcast providing you with the latest perspectives of thought leaders from the accounting and finance profession. If you like what you heard and you'd like to be counted in for more relevant accounting and finance education, visit IMA's website at www.imanet.org.