AI and ethics for the enterprise at KMWorld 2023
While AI paves the way toward new and exciting frontiers, a single component threatens to prevent this technologically bright future from coming to fruition—trust. Without trust, decisions made by AI are pointless and even dangerous, potentially giving rise to both moral and ethical concerns.
Anthony J Rhem, CEO/principal consultant, A. J. Rhem & Associates, and author, Knowledge Management in Practice and Essential Topics in Artificial Intelligence, led KMWorld 2023’s Enterprise AI World session, “AI & Ethics for the Enterprise,” to explore viable and comprehensive ways that enterprises can cultivate truly responsible and trustworthy AI.
The emergence of AI in business has led to innovation in efficiency and automation, ranging in its application from data entry to analysis, customer service, and inventory management. As it continues to grow in the ways it can be utilized, ethics and responsibility will become increasingly crucial, according to Rhem.
“Let’s get ahead from an ethical standpoint,” suggested Rhem. “Let’s get AI ethics baked into what companies are already doing.”
Rhem then defined AI ethics, explaining that it is a collection of moral principles, processes, and techniques intended to advise and direct the development and responsible use of artificial intelligence technology.
To create and nurture responsible AI, Rhem provided a list of requirements that ethical AI should adhere to. Ethical AI must:
- Be trustworthy for businesses and individuals alike
- Allocate responsibility and ensure that there are measures in place to hold the right entities accountable for the consequences of AI decisions
- Create algorithms that are free from biases, ensuring that decisions made by AI are fair and do not discriminate against any individual or group
- Respect user privacy and ensure that data is collected, processed, and stored securely, with consent and in compliance with relevant laws and regulations
- Require rigorous testing, robust security measures, and continuous monitoring to protect against threats and ensure the safety of human users
- Include managing the transition for workers displaced by automation and ensuring that the economic benefits of AI are distributed fairly
- Align AI systems with societal value like dignity, respect, and autonomy
- Require collaboration across borders and sectors, as AI systems often operate on a global scale
“Standards are just the beginning of getting ethical AI under your belt,” cautioned Rhem. “At the end of the day, we have to implement it.”
Implementation begins with data, expressed Rhem. He offered these guidelines for incorporating an ethical AI strategy and application:
- Start with a clear user need and public benefit
- Be aware of relevant legislation and codes of practice
- Use data that is proportionate to the user need
- Understand the limitation of the data
- Use robust data standards, policies, and procedures
- Make your work transparent and be accountable
- Embed data use responsibly
“In an organization, data is king,” said Rhem. “To get ethical AI, you need ethical data.”
He then suggested examining the innumerable AI policies and standards that exist today to discern what you might want as a foundation to produce ethical AI. Adopting a diverse AI development team, a human-centered approach, and leveraging a varied volume of curated, unbiased data can also usher in a comprehensively responsible AI strategy.
“Once you deploy the tool, you’re not done,” concluded Rhem. “It must be constantly monitored and managed to ensure that it runs as intended.”
KMWorld returned to the J.W. Marriott in Washington D.C. on November 6-9, with pre-conference workshops held on November 6.
KMWorld 2023 is a part of a unique program of five co-located conferences, which also includes Enterprise Search & Discovery, Enterprise AI World, Taxonomy Boot Camp, and Text Analytics Forum.