“The ethics curriculum in the department has changed in a variety of ways since I arrived at Rice University in 2019,” said Rodrigo Ferreira, Assistant Teaching Professor of Computer Science.
“Some of those changes have had to do with things happening in the world – and how those have impacted people’s thinking about ethics in general. Throughout the pandemic, people felt very disconnected. We realized our whole lives were mediated by computer technology — work, play, and rest all took place through our computer screens and mobile devices.”
Three years of social and political upheavals also added to the public’s pandemic-induced technology awareness, and students in Ferreira’s computer science ethics classes seemed grow more open to thinking critically about how humans relate to technology.
“These situations have helped us further think about how we relate to one another and to the significance of paying attention to each other. We began to scrutinize the role of computer technology in our society — how to take manage it, and how we can use it to help shape the world in the way we want it to be,” said Ferreira.
He cites a quote often attributed to William Gibson, “The future is already here, it’s just not evenly distributed,” and asks students in his ethics courses what they want for the future. Do they want a future where certain people benefit unevenly from computing advancements, or one where computing helps make the world as a whole more fair, more careful, more compassionate?
“Content in the ethics curriculum evolves with world events, but the curriculum itself is also expanding,” Ferreira said. “I was teaching one undergraduate class in 2019, and the wait list was long. We also found that numerous graduate students were interested in taking an ethics course. So, we reconfigured the curriculum. We now offer one general Computer Ethics course that is twice its original size and available to both undergraduate and graduate students. In addition, we developed new graduate level courses on Data Ethics and AI Ethics.
This semester he is also offering a new Ph.D.-level course on the Ethics of AI and Robotics. It is designed as a practical workshop where students discuss various contemporary issues in ethics and then apply ethical principles to their doctoral research. As a part of this course, Ferreira is currently working with Ph.D. students in the Kavraki Lab on Computational Robotics, AI, and Computational Biomedicine.
“The students are already doing an amazing work in using computational tools to help people, such as developing models that can help cure critical diseases such as cancer,” he said.
“This class builds on top of their existing efforts as the group works together to think through the ethical dimensions of their work. Where does the data for their models come from? Is there an uneven representation of certain social groups in this data? How applicable will the results be for other social groups that might not be as sufficiently represented? We pose and discuss questions like these because the students want to ensure their work can benefit everyone, and not just certain populations.”
Ferreira remarked on how both graduate and undergraduate students are intrigued by questions of fairness in machine learning. It is always the topic that generates the most interaction and prompts the most side projects or additional study.
He said by the time students enroll in a computer science ethics course, most of them are familiar with issues related to algorithmic bias. They’ve read or heard stories about search engines showing certain images over others or how some algorithms tend to favor certain social categories over others in areas such as crime prediction, hiring, and credit lending analysis.
“What the students are usually surprised by is the level of depth required to confront some of the problems. At first, some students tend to look at these issues and think, ‘OK, no problem. If the data going into the algorithm is biased, we will just need to find better or more data to correct it.’
“The students sometimes suggest the problem can be solved by further training models to recognize and mitigate bias in an algorithm. Clearly this can help address the problem to a certain level, yet when we look at the problem of bias in further detail, often the problem isn’t in the data. If the data accurately reflects the world, then the problem is with the world. The algorithm is probably mirroring the same behaviors that humans have been enacting through thousands of individual decisions over time and reinforcing them under the veil of a supposed objectivity.
“So, what do we do?” Ferreira asks his students.
How to confront this deeper issue is what he helps students explore in the Rice CS ethics courses. He has the students look at theories of fairness as well as borrowing concepts from philosophy, history, cultural studies, communication theory, post-colonial theory, and science and technology studies —all with the intent of understanding this deep problem at its core.
It is when Ferreira starts to connect contemporary issues in computing to larger social, cultural, or historical contexts that he sees his students light up from within.
“They realize the issues are not just technical problems with technical answers, but sociotechnical problems that require solutions combining technology with intentionally plotted, socially responsible outcomes,” said Ferreira.
“In my mind, it’s never been the case to teach ethics courses by lecturing students on what is right and wrong. For me, it is about shedding light on cases that interest students and then exploring the social, cultural, and political dynamics of these cases. This also helps students engage with the material at a practical level, so that their future work will be informed by ethical and social considerations from the start.”
Not only have the open dialogues helped Ferreira’s students be better prepared for critical discussions in the future, participating in the dialogues also made them better writers and speakers at an academic and professional level. In addition, multiple students have reported significant shifts in their approach to future projects, employers, and research.
Ferreira said, “After completing my course, one student came back to say they redirected their capstone project to work on a fake news detection algorithm. Another student said she had two separate offers from large technology corporations and that ethics had played a big part in her decision, ultimately choosing to go where she felt that she could make more of a positive social impact.”
He often reminds students that ethics has always been a part of computer science and that its significance will only continue to increase over time. Users care about the social impact of computing; students and soon-to-be developers care about these issues too --so companies are increasingly taking notice and looking to add value in that regard. This creates an opportunity for students in his class to be better prepared to address complex questions for the world ahead.
“It is very satisfying to me when students tell me they have taken what we have learned in class and are actively trying to use computer science to do good in the world.”