“Over the last decade, interest in ethical issues related to computing, especially concerning artificial intelligence (AI) and big data, has skyrocketed.”

Why should philosophers—as teachers, as researchers, as members and administrators of philosophy departments—care about this? And what should they do in response to it?

The following guest post is by members of the Value Analysis in Design team at Northeastern University—John Basl, Katie Creel, Meica Magnani, Vance Ricks, and Yafeng Wang—and the Embedded EthiCS team at Harvard University—Jeff Behrends and Matt Kopec.

In it, they articulate how philosophers and philosophy departments might leverage a particular curricular approach, embedded ethics in computer science (CS), to help themselves and their programs to flourish while also helping to address important challenges in computing (or other fields).

This is the tenth post in the Summer 2025 Guest Post Series at Daily Nous. (As with the previous posts, it will be pinned to the top of site’s homepage for a few days.)

Philosophers and Embedded Ethics

by members of Northeastern’s Value Analysis in Design Team
(John Basl, Katie Creel, Meica Magnani, Vance Ricks, and Yafeng Wang)
and members of Harvard’s Embedded EthiCS Team
(Jeff Behrends and Matt Kopec)

Over the last decade, interest in ethical issues related to computing, especially concerning artificial intelligence (AI) and big data, has skyrocketed. Within philosophy, this has been most obvious in job market trends; demand for specialists in the ethics of AI within philosophy departments has surged and interdisciplinary hires between departments of philosophy and computer science are increasingly common. Perhaps less obvious within philosophy as a discipline is the uptick in a particular approach to ethics curricula within CS and the role that philosophers have been playing in this approach. This approach, often referred to as “embedded ethics,” attempts to integrate ethics education within existing computer science courses. Rather than having a distinct course focused on ethical issues in computing or having a distinct unit within a core computing course dedicated to ethical issues, embedded ethics attempts a tighter marriage of technical and ethical content.

In this post, we want to share a bit more about what embedded ethics is and explain why it’s in the interest of the philosophical community to contribute to embedded ethics programming. While computer scientists and engineers have worked with philosophers in various ways for decades, embedded ethics in its modern form takes inspiration from a program developed at Harvard University to integrate philosophy into computer science. Shortly after Harvard piloted the approach, a number of other programs followed suit through various mechanisms including Mozilla’s Responsible CS Challenge, consultation with Harvard’s team, or independent initiatives at a number of institutions. The tally of successful programs where philosophy plays a central role in embedded ethics programming now includes Northeastern University, Georgetown University, the University of Toronto and Stanford University, among others. More recently, the Mozilla Foundation and the NSF have funded programs that are expanding philosophically-informed embedded ethics programs to a much wider range of institutions, including a large number of minority-serving institutions.

Different embedded ethics programs will describe the goals of their program differently, and there is a lot of variation in how programs implement embedded ethics. Very broadly, embedded ethics programs are oriented towards goals such as making students more critically reflective about the design, use, and governance of computing technologies. A typical implementation of embedded ethics realizes these goals via the development of modules that can be inserted into a course that marries the technical content of the course to ethical issues that topic raises, typically over one or two class periods.

As an example of some interventions, one module developed for an upper-level course at Northeastern on data mining and machine learning asked students to evaluate potential tradeoffs between different philosophical accounts of fairness before connecting this to technical work on tradeoffs between fairness metrics. As another example, one module developed for an advanced-level computer vision course at Harvard examined the implications of deepfakes on the level of trust in our society and how we might develop new technologies to regain the trust in each other that we are quickly losing due to this new technology.

Some embedded ethics programs aim to ensure students encounter a wide variety of modules across courses in the CS curriculum while others use multiple modular interventions throughout a semester within a single course. In order to ensure that modules achieve the aims of embedded ethics, modules are typically co-developed by ethics and CS experts, sometimes in teams. An ethics expert is typically in the classroom to deliver or co-deliver the module, though there are also “train-the-trainers” approaches where the ethics expert helps develop content that is then delivered by someone on the core CS instructional team. In some implementations, like Northeastern’s, students will have taken an introductory course on ethics of technology prior to encountering any modules, and the modular curriculum is developed around values analysis in design (VAD), a central framework that is intended to unify the modules around a specific set of skills, tools, and capacities. Other programs, like Harvard’s, are pluralistic in nature and operate without a core required course.

Embedded ethics programs not only address some significant challenges to responsible computing education generally but they also do so in a way that provides meaningful opportunities for philosophers. For one thing, embedded ethics helps address issues of scale and curricular constraints. The number of computer science majors at your school might outstrip your capacity to teach standalone philosophy courses to those majors for any number of reasons, especially given the importance of ensuring that faculty have the technical competence to relate the content of a tech ethics curriculum to work in computer science. Even if your department does have such capacity, sometimes the required computer science or engineering curriculum is too packed to require students to take another course. The embedded ethics approach can help address both of these issues.

Embedded ethics programming is, we think, especially ripe for contributions from philosophers. Many areas, such as health and medicine, the law, and some sub-areas of engineering (such as aviation) have robust, entrenched ethics ecosystems that serve as a foundation for ethics education, enculturating practitioners through specific norms and policies that govern how they should engage with ethical issues in their day-to-day work. Existing ethics ecosystems can sometimes constrain the role of philosophy in their ethics curricula (especially if these courses are not taught by philosophers), even allowing that philosophical skills could further help students be more reflective, navigate their ethics ecosystem more effectively, and manage issues where existing systems do not provide clear guidance.

In contrast, there is no equivalent ethics ecosystem for most domains of computing. At the same time, there is widespread recognition that the design, development, and deployment of AI systems runs headlong into important, unsettled ethical issues. Many computer scientists we’ve interacted with would appreciate, for example, a procedure to decide which fairness metric to use for a given use case, but they also can see why that kind of procedure isn’t forthcoming! Unless and until much more of computer science develops the kind of ethics ecosystem that has developed in other fields, there is a strong case to be made to computer scientists for the value of philosophically-oriented embedded ethics programming: The sorts of ethical reasoning skills and tools that philosophers are well-positioned to develop are valuable for students in navigating the ethics landscape they will face in their career.

Even if there is room for philosophers to make the case to their computer science colleagues for embedded ethics programming, why might they consider it beyond, perhaps, doing their part to address a broad social need? And, why should philosophy departments work to enable their faculty to contribute to such efforts?

There are a number of potential benefits to developing embedded ethics programming and the kind of close relationship with computer scientists and computer science departments required to develop and deploy such programming. As noted above, there are ongoing funding opportunities to develop and expand embedded ethics programming. These funding opportunities can provide resources for postdocs, course buyouts, and training faculty to better engage with issues in AI ethics. Many of the early philosophically-oriented embedded ethics projects have transitioned from pilots to established programs. For example, based on the success of a pilot with postdocs at Northeastern, computer science and philosophy jointly hired two permanent teaching faculty whose module design and delivery counts in lieu of a standard course within their teaching load. More recently, the philosophy department and the college of engineering partnered to hire another philosopher to help embed ethics content in the engineering curriculum. Harvard’s successful pilot, which started with just four modules designed by a single graduate student, has now grown into an established fixture of the CS curriculum, typically running modules in 24 courses per year. Recently that program has employed as many as four full-time philosophy postdocs conducting research in tech ethics and one permanent staff Program Director and Lecturer in Philosophy.

Embedded ethics programming also helps philosophers form close ties with students and faculty they might otherwise never encounter. This can be a pathway to increased enrollments, majors, or even the development of combined majors. For example, Northeastern has developed a combined CS and Philosophy major that now has more than 200 students. Though not directly attributable to the ethics modules themselves, the resources to hire faculty and develop that major are based on the strong relationship with the college of computer science that was developed through embedded ethics programming. Those relationships with CS faculty in turn can pay dividends for our own research, for example by generating ideas for new research projects or by gaining new insights regarding the technologies over which we theorize. In our experience, those ties can then lead to invitations to join grants alongside computer scientists to do interdisciplinary ethics research, which then brings additional resources and greater recognition to the philosophy program. And perhaps most fundamental are the benefits to the students themselves, whose CS education is enriched by structured opportunities to combine moral imagination and technical ingenuity.

That’s not to say that developing an embedded ethics program is all upside. Creating effective ethics modules with content that speaks directly to work in a given CS course can be hard work. It takes some careful thinking about what kinds of learning goals and philosophical content are appropriate for that context, not to mention developing the necessary interdisciplinary skills and CS content knowledge, and getting student buy-in. The positive news is that new adopters of the model will already have some teaching resources to consult, since many embedded ethics programs, like Harvard, Stanford, and Toronto, have module repositories to get teams started. And due to the other multi-university team projects discussed above, there will be much more to come.

To keep this post manageable, we’ve left out a fair bit about the programs we’ve developed and only highlighted a few of the many embedded ethics programs out there. We invite any readers who have been involved with embedded ethics or closely related programming at their home institution to share their own experiences and to point readers to any available resources they’ve developed. In a future post or in conversation, we can perhaps go into more detail about the challenges and approaches to developing and evaluating embedded ethics programs and content, but we do think that the work is worthwhile and that it is a great opportunity for philosophers to engage students and help develop skills, capacities, and knowledge that many of us see as core to our discipline while having some important impact.

For follow-up inquiries, please contact John or Matt.

art created with ChatGPT and Photoshop