The Power of AI for Good

Artificial intelligence has the potential to drive value across a variety of business sectors, but it is also a powerful tool that can be leveraged to create solutions that have a positive impact on society.

In this video, Ceren Kolsarici, director of the Scotiabank Centre for Customer Analytics and associate professor & Ian R. Friendly Fellow of Marketing at Smith, explores the transformative power of AI in delivering societal benefit. She delves into AI’s role in improving societal outcomes, from enhancing and optimizing health care processes to expanding access to quality education and improving student literacy.

Kolsarici also addresses the obstacles organizations and businesses face when it comes to implementing AI social impact initiatives, such as funding, data quality and access, and lack of infrastructure. She acknowledges the ever-present concerns surrounding safeguarding AI to prevent harm, and believes a balanced regulatory approach is the key.

[MUSIC PLAYING]

Ceren Kolsarici

00:07 What exactly is AI for good?

Let's think about what AI is initially. So, AI is creating human-like intelligence. Thinking of AI, not only as a smart entity, but it's also a compassionate entity. We could consider its impact on improving health care, its impact on improving education, or reducing poverty. So, anything positive we can do for society would fall under AI for good. 

00:34 How can AI improve healthcare?

Think of AI, or machine learning, as a tool to help physicians or health professionals. AI, in general, can scan thousands of images in a matter of seconds. And they are usually, in terms of accuracy and precision, stronger than humans alone. So, that would help us speed up a lot of the diagnostic aspects of healthcare. That would help us optimize treatments across patients. And that would also minimize negative impacts of diagnostics, like false negatives or false positives as well. 

01:20 How is Microsoft using AI to improve student literacy?

They have an AI-powered literacy tool, which they bring to economically under-supported regions of the world, in the United States and Canada, as well as overseas. They bring AI-powered literacy tools to every student. So, ultimately, if we can expand these initiatives, every student, regardless of the zip code, will have access to quality education. 

01:50 What barriers do social impact-focused AI initiatives face?

If you look at social impact-focused AI applications, there are hurdles above and beyond general technological innovation hurdles. The first one is, of course, funding. So, without financial resources, it's impossible to bring these ideas, innovations up to scale to reach enough people in the population. 

The second challenge is infrastructure-related challenges. So, if you think of rural communities, which are consisting of people who need the help the most, those communities, those regions of the world will be the regions that are extremely limited in terms of infrastructure. So, let's say if an AI application is built on robust internet and the region of the world doesn't really rely on robust internet, then that application won't take off in that region. So, infrastructure is a big problem. 

The third, I would say, is access. So, AI can only improve with quality data. However, people who need AI's help the most are people who are hardest to reach globally. So, quality data is also another challenge I would consider with societal AI. 

03:15 How can we safeguard AI to prevent harm?

Canada introduced the AI Security Institute, so it's a way to not only understand the opportunities with AI initiatives, but also understand the risks of the AI initiatives. So, this is one thing we can do at the national level to sort of safeguard the negative consequences of AI. But I think extreme regulation could also stifle innovation, suffocate innovation, and really hurt the development of AI and uptake of technology. So I believe, personally, that there has to be a balance in terms of regulation, government-induced regulation and self-regulation.