Do What We Say, Not What We Do
No, it is not about washing your clothes in a responsible way. Ethics Washing refers to the practice of ethical window dressing. It is where an organization gives lip service to ethics to make it seem as though it acts responsibly but does not do anything to make sure that, in fact, it is occurring in practice. Ethics is mainly for show. In other words, the organization adopts the position: “We are an ethical organization. Just look at these policies.”
An example of ethics washing is the Diversity, Equity, and Inclusion (DEI) policies of organizations. If we reflect on the recent student protests aimed at Jews and Israel, we could argue that universities might have extensive DEI programs for students but when push comes to shove—i.e., blocking Jews from going to classes, hurling discriminatory remarks and so on—the reality of DEI does not match what might be stated in policy statements.
It has become somewhat common to discuss ethics washing in the context of AI. Kapersen and Wallach describe it as “creating a superficially reassuring but illusory sense that ethical issues are being adequately addressed, to justify pressing forward with systems that end up deepening current patterns.”
In this regard, ethics washing is the practice of fabricating or exaggerating a company’s interest in equitable AI systems that work for everyone. An organization that practices it follows a concept that can best be characterized as promoting ethics for the good of all. Some point to Google’s experience in 2019 of creating an Artificial Intelligence (AI) ethics board only to disband it less than two weeks later.
The Importance of Values
When the practices of an organization do not match its stated values, the door is open to ethics washing. The organization might make promises publicly or in the goods and services they provide, but these become misleading or inaccurate claims that harm its reputation. Here is part of the Boeing values statement about engineering excellence:
“A strong engineering foundation enables us to build and maintain our products with safety, quality and integrity in the factory and in service. Our customers expect it. That’s why we will always take the time to get the engineering right.”
Unless you’ve been hiding in a cave these past several years, you know about the safety issues with Boeing’s 737 MAX planes. It appears a week doesn’t go by without another safety issue faced by Boeing.
Oakley and Graham relate ethics washing to greenwashing in stating that ethics washing can be used to describe the practice when organizations paint themselves in a better light than warranted and have no way to back the claims. They identify DEI policies as one example where organizations with DEI policies then discriminate against some candidates in the recruitment process.
The Harvard Business Review points out that “consumers today face a barrage of green-friendly messaging from companies hoping to profit from increased concern over environmental issues. Unfortunately, many of these environmental promises don’t pan out. Research carried out in Europe found that 42% of green claims were exaggerated, false, or deceptive, which points to greenwashing on an industrial scale.
AI Ethics Washing
I have previously blogged about Ethics Washing, relating it to Google’s false steps in creating AI ethics board that flamed out. Beyond that, AI Ethics Washing is a relatively new concept that suggests organizations give lip service or window dressing in claiming it cares about AI principles, but at times it fails to abide by them or even subverts or undercuts AI Ethics approaches. Workplace ethics requires that the actions of an organization match its values and create an ethical organization environment.
Writing for Forbes online, Dr. Lance B. Eliot, a world-renowned expert on AI and Machine Learning, says that in addition to possible reputational harm there are “numerous legal ramifications [that can] bite them and their firm. One is that they didn’t do what they said they did and can be potentially legally held liable for their false claims. Moreover, AI practices might end up violating laws involving societally sensitive areas such as exhibiting undue biases and acting in discriminatory ways.
Firms that exaggerate what they have done with AI might do so because they assume that they will have a sufficient defense if questioned. Absent a comprehensive and cohesive set of metrics, it is difficult, to say the least, to measure compliance. False promises might be made because the organization knows how difficult it is to equate practices with promises.
Rome Call for AI Ethics
Perhaps the least likely place to look for guidance on AI ethics principles is the Vatican, yet in Rome Call for AI Ethics, six primary AI ethics principles are stated as follows.
Transparency. In principle, AI systems must be explainable.
Inclusion. All individuals should be offered the best possible conditions to express themselves and develop.
Responsibility. Those who design and deploy the use of AI must proceed with responsibility and transparency.
Impartiality. Do not create or act according to bias, thus safeguarding fairness and human dignity.
Reliability. AI systems must be able to work reliably.
Security and privacy. AI systems must work securely and respect the privacy of users.
We should add one principle developed by the U.S. Department of Defense in their Ethical Principles for the Use of Artificial Intelligence. The addition is governable, which relates to the responsibilities of the Department (or any organization) to design and engineer AI capabilities to fulfill their intended functions while possessing the ability to detect and avoid unintended consequences, and the ability to disengage or deactivate deployed systems that demonstrate unintended behavior. This principle is essential if the organization is to closely monitor the outcome of AI with respect to its guiding principles and builds confidence in the system.
Conclusion
The practice of ethics washing has been growing in many ways, and the need for ethical principles is more important now than ever before, especially in the AI space. Just think about generative AI and Chat GPT, and so much more that have been evolving over time with little or no stated ethical principles for their use.
Posted by Steven Mintz, Ph.D., aka Ethics Sage, on May 28, 2024. You can sign up for his newsletter and learn more about his activities at: https://www.stevenmintzethics.com/.
