Developing Trustworthy AI
Artificial Intelligence (AI) and the implications this new technology will have in the workplace, the way we do business, and other aspects of our lives have been heavily discussed in recent weeks. In Episode 17 of The Trust Corner podcast, guest Dr. Sebastian Wieczorek, VP of AI Technology and Global Lead of AI Ethics at SAP, shared his thoughts and overall positive outlook about AI.
AI at SAP
A software engineer by trade, Dr. Wieczorek has been studying AI for years, focusing his PhD thesis on the subject then later transferring to innovation management in his career. At SAP, AI is a key pillar of the solutions we provide our customers. “We’re embedding AI capabilities into all the different solutions in the whole portfolio of SAP,” stated Dr. Wieczorek. AI is a central unit that collaborates across different industries and lines of business, and that provides tools and capabilities for development teams across SAP.
Dr. Wieczorek also indicated that SAP took AI very seriously from the beginning—the ethical aspects of AI were discussed the moment AI was introduced to the organization. “We were among the world’s first companies, and I think the first European company to put out ethics guiding principles to basically define what is meaningful,” said Dr. Wieczorek. Since then, SAP became a trustworthy provider of AI by setting structures and best practices, including a steering committee that oversees all the AI-related activities within the company, an external advisory panel that helps the organization make the right decisions, and training made available for employees.
Making AI Trustworthy
A trending topic in AI is how we can prevent bias and discrimination in AI algorithms and ensure that the technology is fair and trustworthy. However, this issue may not have a simple, straightforward solution. “The hard fact is that we cannot fully prevent bias in data, and the discrimination is always going to be a risk,” stated Dr. Wieczorek, similar to how cybersecurity teams can’t promise full and complete security and privacy. The trustworthiness of AI lies in the data and the systems that we use. “We, of course, have to make sure that we have high standards for the data that we’re using, that we’re very thorough in the review processes.” Dr. Wieczorek added that the systems that are hosting these AI algorithms that we have been training need to be constantly monitored for discrimination.
But how can we promote trust in AI? “What we need to make sure is that people are feeling…comfortable, and feeling that they can trust these systems,” stated Dr. Wieczorek. He believes that trust in these systems depends on two things: systems are being developed with good practices and regulation and that they are meeting the expectations of people who use them. People need to see that the system they are interacting with is not harming them and is providing them a beneficial outcome, which will lead them to trust the system more and more.
The Role of AI in the Corporate Environment
Another trending topic has been the possibility of AI replacing certain jobs and processes in the near future. Dr. Wieczorek believes “that many jobs are going to change going forward”, which is a natural tendency in innovation. For example, before cars were invented, there was an entire industry built around horses as a main mode of transportation—feeding, maintenance, carriages, etc. All these jobs were obsolete when the car was introduced to the market. Therefore, “automation and the improvement of the way we’re working is changing the jobs that we’re doing to a large degree.”
Dr. Wieczorek sees this imminent change in the way we work as a positive outcome of AI. “I’m not afraid of [this change],” he stated, “I think what is going to happen is that this efficiency is used by humans in order to focus on other activities…people are more and more pushed into decision-making parts of the work, and less and less on execution.” He then compared AI to an executive assistant that can take care of the tedious tasks most corporate employees deal with, like setting up meetings or sending emails. “In the future, we will all be bosses.”
The potential of AI to change our society is undeniable. However, it is up to us, who develop and train these systems, to uphold ethical practices, follow regulation, and reduce the amount of bias and discrimination in AI.
To learn more about Dr. Wieczorek insights on AI, listen to the full episode here.