On July 28th, I was fortunate to attend a live panel “AI: A World of Possibilities and Perils” with my dear friend and fellow human Jose, Director of Strategic Partnerships, at Pypestream. Before Jose arrived, I joked with a few of the guests around me that this may be one of our last gatherings out in the open before the collapse of society and AI takes control. My joke fell pretty flat and I can’t recommend opening conversations with such grim humor. I’m still recalibrating my social skills here in 2023.
Details of the event described below from UCSC’s event calendar:
AI: A World of Possibilities and Perils UCSC’s Center for Applied Values and Ethics in Advancing Technologies (CAVEAT) at Crown College and the California Arts and Sciences Institute (CASI) present
Artificial Intelligence (AI): A World of Possibilities and Perils, a panel discussion featuring representatives from the arts, sciences, ethics, and economics.
Premiering July 28, 2023, this event marks the first of the series to cover multiple facets of AI’s impact across industries. Live music, complimentary wine and cheese will accent this event at the elegant SandBox venue at 440 Ortiz Ave, #B, in Sand City, CA.
Join us for a lively discussion on the advancements and pressing concerns of Artificial Intelligence. Find out what innovations might revolutionize our lives, and what emerging risks deserve our immediate attention.
Linda MacDonald Glenn, the Founding Director of CAVEAT at Crown College, UCSC, will give a welcome introduction. Edward King, a nationally syndicated radio host, will moderate the distinguished panel. Jennifer Perlmutter, an internationally recognized artist, will represent the Central Coast’s creative community on issues relating to intellectual property and artist livelihoods. Prof. Francois Melese, Emeritus Professor of Economics, will shed light on AI’s costs and benefits, including productivity gain and job disruption. Tom Atchison, an early pioneer in AI development, will provide insights into the various forms of AI technology. Dr. C. Michael Hogan, a leading authority on U.S. energy policy, will examine the potential downsides of AI energy consumption and critical aspects of its accuracy.
A wine and cheese reception with live music will follow the panel, offering an opportunity to meet the panelists.
This exclusive event marks the beginning of an enthralling series, with future presentations featuring experts who will delve into the impact of AI on diverse fields such as filmmaking, journalism, and both nonprofit and for-profit businesses.
Jose and I enjoying an incredible display of human artistry
I was excited to have the opportunity to listen to the panelist for their unique perspectives coming from different disciplines.
One of the distinguished panelists, Dr. Michael Hogan, offered an environmental scientist perspective from using Bard. I found his insights to be particularly interesting so I’ll highlight a few for the twenty or so of you that get this far in the post.
Dr. Hogans’s main message that struck me the most was the need to invest more heavily in the education of the next generation and his advocacy for going beyond the first answer returned from googling or using chatGTP/Bard.
Three takeaways that got me thinking:
One: The consumption of energy from AI exceeds other large consumers of energy like Bitcoin, and Chevron’s worldwide operations. I’m curious about how this was calculated having heard about the below study referenced here. Of course, this is just one earlier model in one company. Even so, there is a big difference in training and usage in terms of energy costs. Thinking about Bitcoin energy consumption beating many countries by this measure, and considering Chevron’s output of a non-renewable energy source I’m a bit surprised by this.
According to Stanford’s Artificial Intelligence Index, it took the equivalent of 502 tons of carbon dioxide emissions to train GPT-3 last year, the model developed by OpenAI that powers ChatGPT, and close to 1,300 megawatt hours of power. As the researchers calculated, this is the equivalent of the lifetime emissions of 8 cars — or 109 cars’ yearly emissions — and enough energy to power an average U.S. home for over 120 years. Of the four models that the report scrutinized, GPT-3 released the most emissions and required the most power consumption.
Two: The claim that autonomous vehicles are 4x less safe than human drivers. My previous understanding was we didn’t have enough data to make conclusions. Do we have enough of a sample size of autonomous vehicles to accurately compare them to human drivers yet?
Three: Glacier ice melt doesn’t contribute as significantly to raising sea levels as water waste. This was pretty surprising to me. I feel like the media I read generally only talks about glacier melt and not water waste.
I hope we get more events like this on the Peninsula. Great to see people come together to discuss and share.