Artificial Intelligence: Its Promise and Problems

Michael Abramson
Dick Duda

Presenters: Richard Duda & Michael Abramson

When: Saturday, August 12th , 2023

Pre-meeting for in-person participants:
6:00 pm – 6:45 pm: (potluck)

Pre-meeting for Zoom participants:
6:30 pm – 6:45 pm: connect to Zoom

6:45 pm – 7:00 pm: check-in for everybody
7:00 pm – 7:45 pm: each presentation is 20 minutes
7:45 pm – 8:40 pm: all-peer discussion

The in-person meeting includes a potluck at the beginning.

Fireside Room at the Unitarian Universalist Church of Palo Alto (for potluck and in-person meeting)
Directions to the meeting:

Our next Humanists at UUCPA potluck meeting is on Artificial intelligence. It’s a hot topic, so this meeting has two presenters: Richard Duda & Michael Abramson, to be followed by a group discussion.

Dick wrote the following for his presentation:

In November last year, OpenAI, a small company in San Francisco, caused a sensation by making publicly available its AI chatbot — called ChatGPT3. ChatGPT’s breadth of knowledge, fluent use of English, and fast response was quite astonishing.

OpenAI and several other companies have used the same general technology — called “generative AI” — to translate languages, write computer code, create images, compose music, and even command robots.

However, if these programs do not know the answer to a question, they simply make something up — in the jargon, they “hallucinate”. Although engineers are working on this problem, it is not easy to solve, which limits their usefulness. In his presentation, Dick will give a basic explanation of how ChaptGPT works, speculate on its possible applications,
and point out the problems with its use.

The following is mostly written by Michael to describe his presentation.

Instead of focusing on chatbots, Michael will look at the bigger picture.
The “landscape of intelligence” includes
– Human Intelligence (HI). Most familiar to us, but not well understood.
– Programmed Intelligence (PI). Drives most of automation around us.
– The specialized Artificial Intelligence (AI). rapidly progressing
– Artificial General Intelligence (AGI). The most powerful, controversial, and dangerous of all.

You will learn that ChatGPT is not an Artificial General Intelligence (AGI), not even close, but it shows what could be an impact of AI having unrestricted access to all knowledge from the Internet.

AGI could be created using ideas borrowed from HI and PI, and can be developed surprisingly fast.

AGI can become an existential threat to humans almost immediately if it gets access to the Internet of Things (IoT). Using less advanced specialized forms of AI as the guardrails against the runaway AGI may or may not work.

Come join us.
For our potluck, please bring a dish of your choice:  vegetarian, salad, meat or dessert. Our meetings usually have 15 attendees.

We usually meet on the second Saturday of each month.

To receive the Humanists at UUCPA monthly announcement and reminder, subscribe to our list: