How to Talk with an Artificial Intelligence w/ Reed Berkowitz

Show Summary:

How can humans determine if an Artificial Intelligence is sentient when sentience is what AI is designed to imitate?

This week, we’re joined by Artificial Intelligence games designer Reed Berkowitz.

Berkowitz discusses former Google Responsible AI team member Blake Lemoine’s claim that Google’s AI chatbot LaMDA is sentient. What made Lemoine believe this? How can we come to understand Lemoine’s belief?

To help us, Berkowitz delivers an AI primer: From AI games to the most popular use cases of the technology, he gives a brief explanation of the terms we need to understand. Then, he discusses an experiment he conducted with an open-source AI chatbot to learn if he can achieve the same results as Lemoine.

Does Lemoine have a case? If not, where did Lemoine go wrong?

About Reed Berkowitz:

Reed Berkowitz is the Director of Curiouser LLC, who has more than 25 years of experience designing, writing, and researching award-winning interactive experiences for companies including Universal Studios, Paramount, Cartoon Network, Peanuts, Sanrio, and many more. He formerly served as the Creative Director of New Games at Latitude, an AI-powered games company.

Follow Reed Berkowitz

Twitter | Website

Show Notes and Resources:

How to talk with an AI: A Deep Dive Into “Is LaMDA Sentient?” - Reed Berkowitz

OpenAI/GPT-3

May be Fired Soon for Doing AI Ethics Work - Blake Lemoine

The Google engineer who thinks the company’s AI has become sentient - The Washington Post

Is LaMDA Sentient? - Blake Lemoine

Hosted on Acast. See acast.com/privacy for more information.