"The Ethics of AI", May 8 (W) at 3:30pm

[Daniel Lake]

 *The Institute for Ethics in Public Life presents "The Ethics of AI," a
hybrid (in-person at the Institute, Hawkins 233, or on Zoom - see below)
forum at 3:30pm on Wednesday, May 8.*

The last few years have seen radical improvements in Artificial
Intelligence technology, including the notable gains in performance related
to the development of new large language models (LLMs) such as OpenAI's GPT
series (e.g. GPT-4). These newer models are far faster and more capable
than earlier AI models, and have triggered a huge surge in interest and
investment in the use of AI for a host of functions.

The particular nature of these AI models and their capabilities is also
raising important ethical issues, due both to their requirements and to
their capabilities. To develop and run these new AIs is very processor
intensive, and this is resulting in major investments in new computer data
centers. One impact of AI-related investment in computing is that the total
energy consumption of global data centers is expected to more than double
between 2022 and 2026, from 460 terawatt-hours (TWh) to more than 1000 TWh.
In addition to making the green energy transition more difficult due to the
increased demand for power, these data centers will both generate a lot of
waste heat and will require a lot of fresh water for cooling. Google's data
centers used 4.3 billion gallons of fresh water in 2021 alone, and the
newest data centers are larger and generally require more water - up to 5
million gallons per day (the same as a town of 50,000 people). The fact
that many of these data centers are located in areas that are prone to
drought just makes the issue worse.

The new AI models also require a lot of data to train on, whether it be
text, images, or sounds. In many cases so far they have been trained on
data publicly available on the internet. While publicly available, much of
the data used is protected by copyright laws and thus its use for training
AIs is questionable ethically, to say nothing of its possible illegality.

Finally, there are ethical issues raised by the use of AIs. The current
generation of AIs are very good at pattern recognition, but it is often
difficult (or impossible) to tell how they came to their conclusions. The
capability to notice patterns that humans will normally miss is extremely
useful in many fields, but the "black box" process by which the AIs operate
and the potential for error raises real ethical issues when these AIs are
used. For example, while we can probably agree that using an AI to screen
medical imagery and identify possible problems for a doctor to follow up on
is good, police or military use of AIs is more problematic. During the war
in Gaza, for example, the Israelis have used a military AI called "The
Gospel" to generate target lists for bombing. While some of the targets
identified through this method were members of Hamas and their bases, it is
also certain (due to the limitations of these programs) that some of the
targets were not valid military targets. While we cannot currently be
certain that Israeli forces attacked any civilians due to their inclusion
in target lists generated by The Gospel,  the very possibility is
problematic.

We are pleased to welcome Dr. Delbert Hart, Chair of the Department of
Computer Science, as a special guest for this discussion. Please join us in
person at the Institute (Hawkins 233) or on Zoom to discuss this important
issue.

Daniel Lake is inviting you to a scheduled Zoom meeting.

Topic: Institute forum on the ethics of AI
Time: May 8, 2024 03:30 PM Eastern Time (US and Canada)

Join Zoom Meeting
https://plattsburgh.zoom.us/j/94248691820
<https://www.google.com/url?q=https://plattsburgh.zoom.us/j/94248691820&sa=D&source=calendar&ust=1714760499565480&usg=AOvVaw3QaHOBs_ecUkScDP0qZe4R>

Meeting ID: 942 4869 1820

-- 

*Daniel Lake*

(pronouns: he/him/his)

Assoc. Prof.  and Chair of Political Science

Director, Institute for Ethics in Public Life

Hawkins 149E

101 Broad Street
<https://maps.google.com/?q=101+Broad+Street+Plattsburgh,+NY+12901&entry=gmail&source=g>

Plattsburgh, NY 12901
<https://maps.google.com/?q=101+Broad+Street+Plattsburgh,+NY+12901&entry=gmail&source=g>

(o) 518-564-5830 <(518)%5645830>

*plattsburgh.edu <http://plattsburgh.edu/>*
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://ls.plattsburgh.edu/pipermail/student-digest/attachments/20240506/7ae455bd/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Flyer for the Ethics of AI.jpg
Type: image/jpeg
Size: 448296 bytes
Desc: not available
URL: <http://ls.plattsburgh.edu/pipermail/student-digest/attachments/20240506/7ae455bd/attachment.jpg>