CTalk: IBM Exec Explains What it Takes to Trust Decisions Made by AI

There are four questions that should be asked before trusting decisions made by artificial intelligence, IBM’s vice president of artificial intelligence technology Aya Soffer said Monday, speaking to Calcalist during the Mind the Tech conference in Tel Aviv.

According to Soffer, these four questions are: Can we trust that the AI’s recommendations are fair? Can we understand why it made the decision? Can we be sure the AI was not tampered with?

And finally, is it accountable?

Working out of IBM’s research and development center in Israel’s northern port city Haifa, Soffer had a role in the development of Watson, IBM’s artificial intelligence-based question-answering computer system.