Data Ethics 101

With the rise of Artificial Intelligence enabling technology, there is a rush to deploy AI enabled solutions in all industries, be it healthcare, security, finance. These solutions either facilitate decision making, or themselves become the decision makers, which leads to serious ethical concerns. Here are some of the key ideas about data ethics that need attention.

Ethics, the system of moral values, is a core trait that distinguishes humans from animals. Acting in ways consistent with what society considers good values is known as Ethical behavior. It demonstrates respect for key moral values that include honesty, fairness, equality, dignity, diversity and individual rights. A society that is not governed by and functions in an ethical manner will not be able to sustain itself.

With the rise of Artificial Intelligence enabling technology, there is a rush to deploy AI enabled solutions in all industries, be it healthcare, security, finance. These solutions either facilitate decision making, or themselves become the decision makers, which leads to serious ethical concerns. Let’s consider a specific use case of facial recognition technology.

Facial recognition technology (FRT) utilizes software to map a person’s facial characteristics and then store the data as a face template. Algorithms or machine learning techniques are applied to a dataset to compare facial images or to find patterns in facial features.

FRT tools that assist with identification, monitoring, and diagnosis are expected to play a prominent role in the future of health care. Consider an individual whose medical insurance rate gets fixed based on the health related information diagnosed based on an image of his/her face. As FRT is increasingly utilized in health care settings, informed consent needs to be obtained not only for collecting and storing patients’ images but also for the specific purposes for which those images might be analyzed by FRT systems. In particular, patients might not be aware that their images could be used to generate additionally clinically relevant information.

FRT tools are increasingly being adopted by law enforcement. Hundreds of U.S. law enforcement agencies are using a face ID service that matches photos against a database of billions of images, the New York Times reported. No user consent was employed by the service in order to use these billion of images of you and me.

There is a dire and immediate need to implement standards of ethics that apply to usage of data. Standards that offers normative guidance related to data, algorithms, and corresponding practices.

Major issues that need to be considered are:

• Validity: is the data/AI model representative of what is being measured?
• Bias and Fairness: is there a systematic skewing of the data collected and/or is there any prejudice or favoritism in the data or model? Are some members of the population being less represented than others?
• Ossification: is the AI model codifying the current state of the world and thereby making it harder to change? Are we building AI models that reflect past mistakes?
• Privacy and Anonymity: does the data reveal the identity of an individual or group of people?
• Ownership: are the rights to the data and related insights derived clearly defined? Is the data acquired with explicit permission? Does the party permitting data usage know exactly how the data will be used? Is there a clear definition of allowed usages of data/AI model?

Data ethics is one of the hotly debated topics with AI practitioners as of now but has not reached a level of concern among community based social justice organizations, whereas the long term impact of it is huge.