As news continues to pour in about the crashes of two Boeing 737 MAX 8 planes, my thinking as engineer turns to how a system of failures resulted in these planes being built and allowed to fly. As with everything, it is more convenient to simplify the situation down to one answer, but this situation has many culprits as many levels: lack of funding to the FAA for proper oversight of the aviation industry, the cut-throat profit-drive of the aircraft duopoly that drove Boeing to avoid full training of the pilots and to make features, including safety features, add-on expenses for the plane, the potential hazards of budget airlines opting for lower configurations of planes, and beyond. Instead, we need to see this as a system of failures at many different levels: technical, economic, and political. However, I want to focus on one level of this system: the lack of rigorous error testing of software.
Ian Bogost of the Atlantic wrote a fantastic piece about this a few years ago in the wake of data security breaches at some of the largest tech companies and about the softening of the term engineer when it comes to software. In his view, engineering as a discipline is inherently to serve the public and engineers must uphold rigorous standards because the entire public depends on their work in building roads and bridges and buildings along with our public utility systems. The field of engineering upholds a strong code of ethics to use the knowledge of engineering for the public good and that the approval of an engineer on a project means that they have ensured their design, be it for a suspension bridge or a water treatment plant, meets the high standards of the code. However, Bogost argues that, when it comes to software, engineers don’t hold up those same standards. Because software can be changed and updated whenever and instantly, the same mindset of ethical standards is harder to apply. If you make a mistake building a bridge, it is extremely difficult and costly to correct it, so you must rigorously ensure correct design on the first time. With software, you can fix a bug instantly and update all systems that use it, so the same pressure does not apply. And the issue with the MAX 8 planes was, after all, a software issue.
As I think about this discussion of the softening of the term engineering in software, I also wonder about the term engineer in the context of biological and biomedical engineering. Like other disciplines of engineering, bioengineers’ designs are technical and complex and the public depends on them for their lives. Think everything from a pacemaker to the newest cell therapies. In healthcare, we have depended on the rigorous clinical trial process overseen by the FDA (which in 2018 licensed more therapies than it had in its history) to ensure that our new therapies hold up to standards of safety and care. In medicine, doctors are guided by the Hippocratic oath and engineers are still guided by the engineering code of ethics. I wonder if the clinical trial process as it stands and these two guiding ethical codes will be sufficient to answer the questions our technology will soon pose: should we embrace germline engineering to eliminate genetic diseases from our population? As therapies get more specific and individualized and more expensive, what right do people have to access these therapies and at what cost?
In my first year of college, we studied the engineering code of ethics. Then, as I entered into my discipline-specific classes, it was never mentioned again. The biomedical engineering department at Purdue required a bioethics course, but not the agricultural and biological engineering department. Many of my friends who are civil engineers are going through the professional engineering licensure process, which includes two rigorous exams and, depending on the state, multiple years of professional experience supervised by a standing professional engineering. As for friends who are not in civil? None are going through the professional licensure process. This seems to me to be a large gap in the training of our next generation of engineers, the ones who will be practicing when we become capable of robustly editing the human genome, of creating increasingly complex AI that we will decreasingly understand the workings of, and face the challenges posed by climate change as our global temperatures tick up degree by degree. If the majority of these engineers will not go through the professional license process, perhaps it is time to rethink or update that process to match where and what people trained as engineers do. And for bioengineering, where we are on the verge of needing to answer a series of complex questions about our direction as a biological species, perhaps we need a code of ethics that combines elements of medicine with elements of engineering, and to make sure that people being trained in the field are also trained to uphold this code.