Ethical change requires more than a billion dollars
It needs deep reflection and Institute-wide transformation
In announcing the College of Computing (CoC), President Reif charged us with a bold mandate to reshape the Institute to deal with the challenges of the day, with two out of the five tasks dealing with ethical usage of technology, both in education and in research. With this bold ethical mandate, and all the similar ones to follow, I find hope in a better MIT. We might one day consider nefarious impacts of research before we start on a project. We might be willing to reject research funding that compromises our values. Our future students might leave MIT with a deep concern for the impact their work makes on the world. And I might be proud to belong to such an upstanding university.
At the same time, I remain skeptical. The announcement and direction of the college appears very top-down in its approach, and I fear that lack of buy-in from faculty, staff, and students will prevent the CoC from creating the transformative change necessary to fully embrace the ethical challenges mandated by President Reif.
Now that the CoC celebrations are wrapping up, we should recognize that convincing MIT faculty, staff, and students to value acting ethically will require much more than a few new professors and a billion dollars. We need to take a long, vulnerable look at ourselves and reassess our priorities. We likely need to set aside other core values, like faculty autonomy and maximizing endowment growth, so that MIT can better serve its mission. This will be difficult, but it is necessary to truly reshape MIT.
During my past four years at the Institute, I have experienced pervasive ethical avoidance and failures, which highlight how multifaceted and daunting President Reif’s mandate is. A year after arriving on campus, I saw how the administration turned a student movement for MIT to reject companies that fuel climate science denial groups into Low Carbon Research Centers, which accepted money from at least one company that at the time belonged to a lobbying group involved in climate science denial, growing MIT’s research budget at the potential expense of perpetuating falsehoods. With regards to faculty, I remember arriving to class with “Black Lives Matter” written across the top of the chalkboard and wondering whether the professor would at least acknowledge the words staring at all of us. He never did. And I remember the numerous times classmates proclaimed that they didn’t care about what they worked on, or the impact it had, as long as it was challenging. This mentality mirrors the ethos of hackathons, which often condenses complex societal issues into small tech-based problems alleviating us of the trouble of asking larger, and often ethical, questions. For example, in a 2015 Clean Earth Hackathon I attended, we were asked to develop a technology that better sorts plastic fragments out of broken glass for a recycling sorting facility, ignoring larger questions like why we purchase and discard so much plastic and glass.
We can prevent these types of mistakes in the future by thinking about how ethics applies to our work. The administration can reassess when MIT should not accept funds which would contribute to misinformation and perpetuate inequalities within society. Some of these changes may infringe upon other Institute values, which can create political roadblocks, and the administration needs to be ready and willing to address the concerns created by these shifting values. For example, creating an ethical review board which evaluates broader impacts of corporate-funded research projects might go against the longstanding value of faculty autonomy by restricting faculty members’ ability to accept research money, so the administration would need to roll this out in a way that reduces its perceived impact on faculty autonomy. Thus, the administration should evaluate how to tackle these issues before informing the entire faculty. To ensure that changes to the required curriculum address Reif’s ethics mandates, the administration will need to develop metrics that measure the necessary ethical changes and collect data before and after that curriculum change to determine if the change met the requirement.
Faculty need to tackle their own discomfort with ethical issues and bring their newfound ethical framework into the classrooms and research labs. When motivating students with a real world example, they should resist painting the story as entirely good and without any ethical questions. They should integrate ethical questions into assignments when using realistic problem statements and projects. When accepting corporate research grants, faculty should make sure their research does not become public relations material for companies, especially if the research could be misused to prevent certain benefits to society, like fixing the healthcare system or tackling climate change.
As students, we must grapple with the knowledge that our work has the potential to harm society, and that we are responsible for the impacts of technology that we help bring into this world. In our CI-Ms, at conferences, or in publications, we should mention any potential ethical concerns alongside any societal benefits, and when attending others’ talks, we can ask questions if they forget to mention ethical concerns about their new technology.
None of this is easy or natural, and the parties over the last few days have likely done little to help us question our moral shortcomings, but this is precisely what we must do. If, as MIT’s administrators, students, and faculty, we are not serious enough about ethics to consider it in line with our other priorities, the CoC could perpetuate, or even exacerbate, the use of technology and AI in unethical ways to manipulate politics, increase economic inequality, and amplify extremist views. If this is the case, then the ethical promises MIT has made with the CoC will become the fake news that our technology is so successful at spreading.
Mark Goldman is a graduate student in the MIT Department of Chemical Engineering.