Senior Vice President, Swiss Re
Jerry Gupta is an executive with experience leading data science and technology transformation initiatives and programs and managing both technical and business teams. He is an SVP at Swiss Re, one of the world’s largest reinsurance companies.
Previously, Gupta was the Global Head of Program Management at Amazon. Before that, he helped launch the innovation and venture groups at Liberty Mutual Insurance.
He has launched new businesses both as an entrepreneur and within Fortune 100 setting, conducted due diligence on transactions worth more than $3 billion USD in aggregate value and raised more than $25 million USD in private placements. He has also been on the board of, or an advisor to, several start-ups in the US and in Spain.
Gupta has an MBA from MIT Sloan School of Management and a master’s degree in predictive analytics from Northwestern University. He has also completed his MS in computer information systems from Bentley University.
Getting to know...
Jerry, in a recent report we published, we found 67% of enterprises don’t monitor their AI models to ensure their continued accuracy and prevent model drift, what are your thoughts about this statistic?
This number is not too surprising, if anything it might be too low. Model maintenance is a non-trivial activity, and as a general rule of thumb, takes up to of 33% of the time and effort it took to build the model. One has to not just look at the changes in input data that results in the drift, but there is an “operational drift” component to it as well because there is a tendency of business and data teams to disengage once the model has been deployed. Most organizations face severe resource constraint when it comes to data science resources and operate in an environment where the project backlog far exceeds the available resources. In this scenario, priority is almost always given to initiating new projects as opposed to revisiting prior work resulting in chronic under-investment in maintenance activities. Unfortunately this “deferred maintenance” can really hurt the company. Model drift can lead to risks that range from poor decision-making to more serious issues that can carry regulatory and reputational risk. Given the demand-supply gap in data science field, it is unrealistic to expect that most companies will be able to hire talent to fulfill their ongoing needs. Therefore I am a big advocate of re-training existing employees so that they are able to take on a bigger share of the data science workload, while leaving only the complex problems to experts. In fact, I have developed and launched data science training in my company in order to demystify data science and enable routine activities to be undertaken by business teams.
Why did you want to be part of the Business of Data Advisory Board?
Data science is one of the most important functional areas but there is a great deal of misunderstanding about this subject. Topics like ethics, privacy etc. are still not well understood. My goal would be to use this platform to promote the concepts of ethical AI so that we can harness its power in a safe and socially beneficial manner. AI will be the defining technology of this century and if we do not take a holistic view of its development, application, benefits, and risks, it can pose serious risk to humans. I lean towards the view that it’s benefits outweigh the risks, but we need to be fully cognizant and ever vigilant. I think it is unfortunate that we have been so focused on the automation and customer experience related benefits of AI, that we have not paid as much attention to its second-order detrimental effects. From an economic point of view, we have to really worry about its tendency to create natural monopolies. But the human impact of continued automation cannot be under-estimated. Humans are resilient and have overcome similar disruptions in the past, and it is highly likely they will use AI to augment human capabilities, but there is a big enough likelihood of an adverse outcome, requiring caution on our side. Over the past few years, I have been a vocal advocate for code of ethics for Data Scientists and development of standards regarding AI development and usage. My hope is that this platform will bring together like-minded people who believe in ethical AI so that we can work towards a practical and humane solution to this potential problem.