Helping Mold Ethical AI
A new MSAI course coming this spring will focus on the risks and moral issues behind the development of artificial intelligence, other modern day technology and how today’s students can work to create a brighter tech-assisted future.
A new course in Northwestern Engineering's Master of Science in Artificial Intelligence (MSAI) program aims to discuss potential risks coming from emerging technologies and teach students the importance of ethical AI development.
The course, which will meet for the first time this spring, is called Emerging Risks for Humans Interacting with Technologies and will be taught by Oleg Evdokimov, group product design manager at legal-technology company Relativity.
Evdokimov is currently teaching MSAI’s Human-Computer Interactions (HCI) class for the third time. The new offering is a logical extension of the current course, he said.
Whereas the existing HCI course aims to emulate the modern day product development team-based approach to creating new digital products that serve humans, the new one will discuss how we all are collectively creating our technologically infused future. The volumes of digital data we provide with every interaction with connected technology – be that a website, our vehicles, or even internet-enabled kitchen devices – are literally shaping the growth of AI while reshaping our daily lives.
“Even though you're consuming certain applications or websites or interacting with your refrigerator, you're also at the same time a builder of those applications because the data is being used for a variety of purposes. We are becoming creators of the products by supplying the data to improve them, while also becoming a product from the perspective of those who want to use our data in a variety of ways” Evdokimov said. “Those roles are somewhat blurrier now, and I think there's a much bigger impact of regular people on the field as opposed to earlier in HCI’s history.”
The new course is designed to help turn MSAI students into increasingly aware ethical developers of AI products that focus on the betterment of the people who use them.
Given the emerging nature of the technologies, a key component of the new course will be conversations about the implications of the technology. While doing that, students will get hands-on experience building and then adjusting simple recommendation engines to experience the effects of small modulations. With that understanding, they can then help identify how to solve the problems caused by the changes.
“I want to teach the class to spend just as much, if not more time defining and understanding the problem, because if you have the best solution to a wrong problem, it doesn't really matter,” he said. “What’s important for students to learn is to spend time embracing those affected by that problem.”
Evdokimov has a wealth of experience doing that. He has been involved professionally in user experience design for more than a decade. In his current role at Relativity, he manages a team of designers and researchers focused on improving the company’s legal technology software.
He said his focus has always been on relationships – between humans and other humans, as well as between humans and technology.
“The way I define it is I enjoy finding ways for groups of people to work better together,” said Evdokimov, who graduated from Northwestern Engineering’s Master of Product Design and Development Management (MPD2) program in 2017.
Technology should be assisting, not hurting, those relationships, but Evdokimov said his course will point out examples where harm has happened – such as social media networks that led to shorter attention spans and weaker community connections or businesses that sell user data that many don’t realize is even being collected.
Those examples underscore the need for careful attention to the ethics of human-computer interactions, Evdokimov said. The impact of these interactions spans from individuals, small groups, and businesses, all the way up to nations, regions, and, ultimately, the global community.
“The general structure of the class is to go through the individual all the way to the planetary level and analyze and discuss how AI and human computer interaction is affecting each layer,” Evdokimov said. “I'm very excited about this course.”