Opinion guest column

College of Computing working groups are missing experts on ethics

The college’s task force has made no measurable commitment to ensuring the ethical development of technology

When President Reif announced the College of Computing (CoC) over four months ago, he claimed that one of its three major aims was to promote the ethical usage of AI and its societal impact. “As computing reshapes our world,” Reif said to MIT News, “MIT intends to help make sure it does so for the good of all.” With recent uses of AI to manipulate elections, polarize our society, and exacerbate racial pricing discrimination, I completely agree with the need to more strongly embed ethics into the curriculum and research at MIT. What disturbs me is that the strong language of these announcements, like “the need for bold action, at scale and with speed,” appears to contradict what I hear from graduate students in EECS whose department and/or advisors assure them nothing will change. No new ethics course requirements. No additional oversight on the societal impacts of their research. Nothing.

As I wondered whether the Institute was actually serious about ethical change, Provost Schmidt announced the creation of five working groups to guide different parts of the CoC. As I read his letter, it seemed to eat away at my hope for an ethical MIT. While ethics was the focus of one of the working groups, it did not appear to be part of the other four. I felt MIT had isolated ethics from the discussions happening in the other working groups, like faculty assignments, computing resources, and curriculum development.

But as I asked myself, “Who is not at the table?,” my cynicism grew. Of the 100+ committee members listed, not one comes from the Office of Religious, Spiritual, and Ethical Life, which recently had its name changed to emphasize its role as an ethical resource and provides support for individuals and the entire Institute. During her inauguration last year as the Chaplain of the Institute, Reverend Kirstin Boswell-Ford said, “We are here to help guide the MIT community toward being more inclusive, more caring, more supportive, and more ethically minded.” Excluding experts who grapple with difficult, ethical questions through their daily work continues the trend of undervaluing religious and ethical voices at MIT, such that the committee will lack the ingenuity to reshape MIT.

Also entirely missing from the room are researchers from the MIT’s Algorithmic Justice League, a group widely recognized for its work on bias within AI. Since the announcement of the CoC, the group has been mentioned in Time, Bloomberg News, and DailyMail among other outlets. Excluding the most widely known MIT researchers working on ethics and AI from the working groups makes me question the ethical commitments stated when publicising the CoC. Given both the segregated and exclusionary nature of the working groups, "death by committee" feels like the most probable fate of ethics at MIT.

Despite my distrust, I asked the Provost about commitments the steering committee has made to ensure that ethics will not be swept aside as a billion dollars gets sliced up and fought over by various schools and departments. In his response in an email to The Tech, Schmidt reinforced the importance of ethics: “It’s also incumbent on us to conduct research in a way that not only advances the technology, but also advances our collective understanding of how to deploy the technology. Wherever the research leads, there should be policy work in parallel. And even beyond considering policy and the regulatory framework, we should approach any given research problem and solution design with an ethical lens.” This statement reassured me that I had not been misattributing the large ethical promises MIT made when launching the CoC, but did not indicate any measurable commitment to ethics.

When answering a question about the act of separating ethics into its own working group, Schmidt mentioned that the co-chairs of the working groups would have regular meetings, so that ethical considerations can be raised to other committees. This idea sounded nice, but it still doesn't ensure ethics is truly integrated into the other meetings.

While the Provost explained his general views on ethics, he did not mention any specific metrics that would ensure that the CoC will embrace ethics. Has any fraction of the billion dollars been earmarked for the ethical usage of AI? How many of the 50 faculty hires will be reserved for experts on the ethical usage of technology? To these questions he responded, “Answers ... will emerge from the activities of the working groups.” Given who is excluded from the working groups and how they are structured, his message amplified my anxiety.

Despite all this, I still have some hope. If we speak out publicly, MIT has to listen, just like it has in the past. For example, after people started publicly discussing issues with finding support after experiencing sexual violence at MIT (The Tech published four such articles in Spring 2014), MIT conducted a survey on sexual assault in late 2014, created the Title IX office in 2015, increased transparency regarding investigations in an annual Title IX report, expanded prevention and support programs through Violence Prevention and Response, and required MIT faculty to complete online gender-based violence training.

Just like the #MeToo movement provided a critical mass of survivors with the confidence to drive change, the CoC provides us with an opening to demand MIT act ethically. Investing your time in this endeavor could not be more valuable. Ensuring MIT fulfills its commitment to ethics over the next year will impact humanity for generations. Take a moment, think of a way MIT could be more ethical, and share it on the idea bank. If administrators don't promise this semester that they will do what you suggest, ask them why. Then refine your understanding of the issue. Then ask them again. If they still don’t listen, call them out publicly. And continue to ask until MIT commits to the ethical promises they made when announcing the CoC.

Tell MIT what it needs to do to become ethical (and support what others suggest) at comptf.mit.edu.

Mark Goldman is a graduate student in the MIT Department of Chemical Engineering.