View all topics
Back to CompTIA

Operationalizing Ethical AI: From The C-Suite To The Data Pipeline

  • 2 wk ago

Since the pursuit of machine learning began in the mid-20th century, the technology industry has focused on building artificial intelligence (AI) capabilities that replicate human intelligence. It’s only in the last five to 10 years, as AI has become more of a practical reality, that discussions around ethical AI have reached the mainstream. And although there’s general agreement on the principles of ethical AI (e.g., transparency, justice and fairness, non-maleficence, responsibility and privacy), there’s little agreement on how to apply and operationalize them in an organization.

Consider findings from a recent survey by IBM, which show that despite a "strong imperative" for the need to advance ethical AI, there's still a gap between business leaders’ intentions and meaningful action. Nearly 80% of CEOs stand ready to embed AI ethics into their companies' business practices, but less than a quarter have operationalized them. And less than 20% of those surveyed said their company’s actions were consistent with its AI ethics principles.

Sadly, these findings are neither shocking nor uncommon.

The Barriers To Successful Implementation Of Ethical AI

Researchers from Microsoft Research and Carnegie Mellon University worked with nearly 50 machine learning practitioners from more than a dozen tech companies to compile an ethical AI checklist. In the course of their work, staff members heard a common refrain: Speaking up about ethics issues exacted a social cost and could adversely impact a person’s career because advocating for AI fairness could slow the pace of work and lead to missed deadlines.

When ethical issues do arise, all too often the knee-jerk response for most companies has been to implement better algorithms or tech-based controls that help rein in bias or other unethical practices. But most of these same practitioners told the researchers that any solution to ethical AI issues should be both technical and non-technical in nature. Why? Reasons included:

Continue reading:

Reply Oldest first
  • Oldest first
  • Newest first
  • Active threads
  • Popular
Like1 Follow
  • 2 wk agoLast active
  • 1Views
  • 1 Following
Powered by Forumbee


View all topics