Seven AI-Based Ethical Issues for Project Managers

 

Applying AI algorithms to make project decisions or using generative AI to resolve project issues can create ethical concerns for project managers. Organizations typically have data privacy and security policies, and governments have privacy regulations to protect personal data. Using AI technology has additional ethical requirements for project managers, and seven of these are reviewed below.      

Informed consent. Typically, informed consent is the right of an individual to provide knowledgeable agreement to organizations that want to use their personal data. Using AI technology involves a new perspective on this requirement. Without informed consent, there is a liability when resources are identified in a resource plan or listed in project scheduling software, and the data is shared across organizations or with contractors. This example might fall under data privacy policies, but the possibility of sharing data without consent is more significant when using AI tools. AI algorithms can analyze and provide insight into efficiency or inefficiency for named resources, which may not have been included in informed consent. The analysis and output may require more vigilance.

Bias in the data. Historical data is known to have bias. For example, the bias can be against a specific gender, ethnic background, or age. AI tools are used for resource allocation and capture data on resource efficiencies. How is bias removed from the process?

Corrupt data. There is an adage that states, “Garbage in = garbage out.” From an ethical perspective, project managers must evaluate if decisions are made based on bad data.

Lack of maintenance. This concept is described well in the book Weapons of Math Destruction (O’Neill, 2017). Data used for AI algorithms needs to be updated regularly. Would you ride in an elevator that has not been serviced in 30 years?

Poor interpretation. Project managers who use AI need a basic ability to understand statistics and how they influence the results. For example, should a data point that is an outlier be ignored, or is it the start of a trend? Mindlessly implementing AI-generated results can deliver poor outcomes. Project managers have a personal ethical responsibility when using AI results instead of blaming the tools used. 

Inaccurate results. AI-based algorithms can generate inaccurate results. Knowingly making decisions based on erroneous output is inappropriate. Taking action without realizing the results are inaccurate means the organization has failed to take responsibility for proper training.

Untraceable algorithm. Some large algorithms do not provide insight into how they arrived at the results. This has created a new field of knowledge known as Explainable AI. There are methods and practices that can be implemented for humans to provide oversight so the reasoning or logic behind algorithm results is understood.

Accountability

As outlined below, organizations must provide the framework for project managers to properly assess and address ethical issues due to AI technology.

1) Ethics compliance. These are policies and procedures for how AI is deployed and managed within the organization. They need to address the issues and provide direction for project managers. They define how to avoid ethical problems and manage them when they occur.

2) Ethics governance. A person or group with a higher-level perspective can monitor and ensure policy adherence. This oversight becomes a source of knowledge and support for clarifying or identifying gaps and omissions.

3) Training. The most important component is to provide training with examples for project managers to understand how to manage ethics in an environment that is becoming increasingly filled with AI-based tools.