Researchers Seek to Teach AI Programmers Ethics Practices
Microsoft researchers are leading an effort to create an AI ethics checklist for companies to avoid thorny issues of discrimination in programming affecting women, minorities and others, reported in machine learning products.
A story on venturebeat.com describes how Microsoft interviewed 50 engineers from more than a dozen companies for a paper to help solve the problem of flawed ethics in AI.
“Altogether contributors to the checklist are working on 37 separate products, services, or consulting engagements in industries like government services, finance, health care, and education,” according to the Venturebeat story. “Interview participants were not identified by name, but a large number work in AI subfields like computer vision, natural language processing, and predictive analytics.”
Andrew Ng, founder of Google Brain, and Landing.ai, described at the NeurIPS conference how his company and others are participating in a Microsoft led effort to develop ethics standards for their projects. Ng said the engineers need more guidance that a list of dos and don’ts on how to implement the policies.
Accordingly, the project involves explaining how the check list can align with existing workflows, using “six stages of the AI design and deployment lifecycle rather than on a standalone set of ethics principles.”
Engineers gave researchers in-depth feedback on their struggles and how they have adapted, providing valuable insight for their colleagues through the Microsoft paper, which earned recognition as the best research paper at the ACM CHI Conference on Human Factors in Computing Systems.
“Participants generally said they do not want AI fairness checklists to champion technosolutionism or overprescribe technical solutions to social problems; that solutions must be customizable to adapt to different circumstances; and must take into account things like operational processes and company culture.”
Implementing an AI ethics checklist is more complicated than most companies are equipped to handle, leading to the hiring of AI ethics officers and other professionals to work full-time on such efforts, according to a story on futurumresearch.com.
read more at futurumresearch.com