Tuesday, June 12, 2018

The Need for the Human Factor in an AI World - Tip #191


How do you imagine a future with AI? Will the world be run by machines and computers? Or are humans still an important part of the picture?

An AI Named Quixote

Quixote is an AI system that learns about ethics and human values by listening to simple stories. According to Mark O. Riedl and Brent Harrison, creators of Quixote and researchers at the School of Interactive Computing at Georgia Institute of Technology, the AI system reverse engineers’ human values through stories; narratives tell Quixote how humans interact with each other and informs the system what the accepted behaviors are in a society. So, it knows for instance, that stealing is wrong.

The innovation may “wow” you, but behind the scene is a person deciding which stories to feed the AI system. H. James Wilson, Paul R. Daugherty, and Nicola Morini-Bianzino believe this and the role of monitoring and making sure Quixote runs properly will be the critical responsibility of a human ethics compliance manager


So yes, humans will still be important in a world dependent on AI. In training and development, in particular, human trainers will play an important part in shaping employees to take on novel jobs.

The Role of Human Trainers in an AI World


In the world of AI, jobs will be vastly different than what we know today. An Accenture study and Wilson, et al. note that these future jobs will be “new,” “unique,” and “novel, requiring skills and training that have no precedents.”

These new novel jobs fall under three categories, according to Wilson et al.:

  • Trainers (e.g., empathy trainer) will be responsible for teaching AI systems on how they should perform. For instance, a trainer makes sure chatbots can detect when a customer uses sarcasm in their communication and respond accordingly.
  • Explainers (e.g., algorithm forensic analysts) are those who can explain to business leaders, how complex algorithms work. They will essentially become the “bridges” for non-technical professionals to understand how AI works.
  • Sustainers (e.g., ethics compliance manager) will ensure that AI systems work as they were meant to work and that any diversion from that function and the consequences arising therefrom will be immediately addressed.
In short, human workers will make sure that AI tasks are “fair, transparent, and auditable.”



Over to you: How else do you see the roles and responsibilities of training and development professionals change in an AI world? Share your thoughts with me in the comments section below.

References

What is the Quixote AI System? (Long version). YouTube/Entertainment Intelligence. March 1, 2016
Brent Harrison and Mark O. Riedl. Towards Learning From Stories: An Approach to Interactive Machine Learning. Association for the Advancement of Artificial Intelligence, 2015
H. James Wilson, Paul R. Daugherty, and Nicola Morini-Bianzino. The Jobs That Artificial Intelligence Will Create. MIT Sloan Management Review Magazine: Summer 2017 Issue
Adi Gaskell. Do We Need To Set Aside Time For Learning At Work? Forbes, July 20, 2017




Ray Jimenez, PhD
Vignettes Learning
"Helping Learners Learn Their Way"

1 comment:

  1. Thank you for defining the three categories, which I find very interesting. Another future role I see is building the Semantic Web. Educators who define and input meaning will be in a critical, foundational position to shape thinking and there for education. Ultimately, all these positions will be limited to the ethical and critical thinking qualities of people, so we cannot blame machines for any negative consequences.

    ReplyDelete

Welcome! Sharing your comments is very valuable learning experience for me and others. Thanks!