The Interaction Hour

Over the past half a decade or so, the field of deep learning has exploded. Artificial intelligence and machine learning are experiencing their moments in the spotlight – both in research corners and in popular culture. As with anything being given so much attention and hype, it’s difficult to separate the fact from the fiction. Users want to understand AI – how does it work? What is it capable of? Why does it come to one decision over another? A lot of effort has gone into opening the proverbial black box to better understand the AI. Not as much time has been spent on the other half of the equation: Humans. Today, we’ll speak to one Georgia Tech researcher, whose team will present work on this primary topic at the upcoming ACM Conference on Human Factors in Computing Systems. What role do humans play in the infrastructure of AI? How can a better understanding of this role lead to better design? For a successful human-AI relationship, do we need to gain trust in our technological counterparts – or do we already trust them too much?

Featuring: Georgia Tech Ph.D. student Upol Ehsan

After more than 15 years as a researcher, educator, and administrator at Georgia Tech, School of Interactive Computing Chair (and Interaction Hour host) Ayanna Howard is moving on to The Ohio State University. There, she will become the university's dean of engineering, a role she says will help her to continue her mission of improving equity and inclusivity for the fields of engineering and computer science. In this episode, we take a look back with Dr. Howard over her time in Atlanta, what the future holds for her, and what we can do to continue to make the world a better place through our research and academics.

For nearly a year, we’ve seen the COVID-19 numbers. It’s been an ever-growing climb on news channels, positive cases and daily death totals, fatality percentages, and more. The constant flow of information has been as overwhelming as it has heartbreaking. But what does all of this information mean, and what do we do with it? How do we know what the numbers are telling us? Clio Andris, whose work on a visualization tool released last summer has been met with much media attention and public use, joins the Interaction Hour to discuss.

The Covid-19 pandemic has necessitated a new approach to education, forcing some students to mix in-person learning with remote and others to learn only from the confines of their own homes. The pandemic has brought to the fore new challenges and potential solutions to address the pressing needs of students and educators, and it could expedite a transformation in how we think about education in the long term. Today, we’re joined by David Joyner, the executive director of online education in Georgia Tech’s College of Computing. He’ll help us understand the pressing needs during the pandemic, the long-term benefits of online education, and how we might bridge the gap between the best of both worlds.

One of the biggest challenges to achieving equity, diversity, and social impact in computing is how we engage with traditionally underrepresented populations in the field. Associate Professor Betsy DiSalvo is the principal investigator on the DataWorks project, a program that has brought employment and engagement to non-data scientists. In this episode, we’ll discuss the accomplishments of DataWorks, explore how it engages those without a background in computational thinking, and how it improves our pursuit of equity in computing.

It is a uniquely challenging time in human history. We are facing a deadly global pandemic that has killed more than half a million people and caused a tidal effect across all levels of our society – from health care to the economy to education to our social lives, and much, much more. In our own country, we are facing civil unrest as we reckon with the impacts of centuries of oppression and disproportionate treatment and opportunity of the Black community. And we must meet these pressing needs in the midst of a technological revolution that, if not handled properly, will continue to compound our past failures and threaten to leave us behind, unable to keep up. In times like these, it is important to remember what – or more accurately, whom – is at the center of all we do in Georgia Tech’s School of Interactive Computing.

Data ethics is central to all we do as researchers, educators, innovators, entrepreneurs, and even consumers. And yet it's an area that is woefully understood by the general public. Companies misuse data and take advantage of consumers who aren't entirely sure what they agree to when they sign the terms of use. Data that was meant to solve one important challenge is repurposed in unintended ways. How can we encourage our researchers to begin thinking about these implications? How can we make data use and ethics more accessible to the general public? All this and more is discussed on this LIVE edition of the Interaction Hour with Ayanna Howard, featuring guests Amy Bruckman, Ben Shapiro, and Amanda Meng.

Are humans too willing to transfer trust to AI systems that may or may not have earned it yet? What factors lead to that trust? What’s the threshold for how trustworthy a system, like autonomous vehicles, must be before we deploy worldwide, and how do we get there?

In a previous episode of the Interaction Hour, we discussed one potential space that could benefit from virtual reality. A group that included one of our faculty, Neha Kumar, was using the technology in the educational space, working with local teachers to develop virtual lessons that showed improved engagement and performance. Today, we return to the topic. Virtual and augmented reality continue to be among the most promising technologies, but what they are, what they will become, and where we will benefit is still up for debate. Even more pressing are the potential pitfalls – like privacy – which, without proper vigilance, could be exploited in much the same ways as social media.

Think about the most recent news headline you read. Was it completely objective, void of any presupposition of truth or language that may lead readers down one particular path of understanding? Or did it, more likely, contain subtle cues about how the message was being framed, casting doubt on its veracity or reliability. Every day, we are inundated with these types of texts that, on the surface, proclaim to be arbiters of truth but, due to simple word choice and message framing, can bias their consumers. Luckily, new tools are being developed to help us become more critical recipients of media. In this podcast, we chat with Diyi Yang about how artificial intelligence can help us identify this subjective bias in text – and how AI itself can reflect our own preexisting biases.