Balancing AI Power with Human Judgment
Key Points
- AI’s potential goes far beyond large language models, yet many people mistakenly treat LLMs as the whole of artificial intelligence.
- While LLMs excel at summarizing and synthesizing information, over‑reliance on them risks “dumbing down” human sense‑making and eroding our ability to develop judgment.
- Our research program focuses on finding a balance between AI assistance and human reasoning, including creating audit tools to safeguard against loss of human capability.
- Using LLMs for primary data collection can stifle imagination, abductive thinking, and the discovery of rare or out‑of‑pattern insights that only independent searching can reveal.
- To preserve essential skills we must keep manual search, synthesis, and representation practices alive—much like climbing with a GPS, we need both technology and personal effort.
Sections
- Balancing AI Potential and Human Judgment - The speaker emphasizes AI's broader capabilities beyond LLMs, illustrates how LLMs excel at summarizing content, warns that over‑reliance could diminish human sense‑making and judgment, and outlines a research program and audit tools aimed at preserving a balanced partnership between AI assistance and human decision‑making.
- Preserving Human Skills Amid AI - A speaker warns that overreliance on language models can erode essential human abilities, urging the maintenance of those skills to navigate uncertainty.
Full Transcript
# Balancing AI Power with Human Judgment **Source:** [https://www.youtube.com/watch?v=dvDGkP3V9Z8](https://www.youtube.com/watch?v=dvDGkP3V9Z8) **Duration:** 00:03:24 ## Summary - AI’s potential goes far beyond large language models, yet many people mistakenly treat LLMs as the whole of artificial intelligence. - While LLMs excel at summarizing and synthesizing information, over‑reliance on them risks “dumbing down” human sense‑making and eroding our ability to develop judgment. - Our research program focuses on finding a balance between AI assistance and human reasoning, including creating audit tools to safeguard against loss of human capability. - Using LLMs for primary data collection can stifle imagination, abductive thinking, and the discovery of rare or out‑of‑pattern insights that only independent searching can reveal. - To preserve essential skills we must keep manual search, synthesis, and representation practices alive—much like climbing with a GPS, we need both technology and personal effort. ## Sections - [00:00:00](https://www.youtube.com/watch?v=dvDGkP3V9Z8&t=0s) **Balancing AI Potential and Human Judgment** - The speaker emphasizes AI's broader capabilities beyond LLMs, illustrates how LLMs excel at summarizing content, warns that over‑reliance could diminish human sense‑making and judgment, and outlines a research program and audit tools aimed at preserving a balanced partnership between AI assistance and human decision‑making. - [00:03:08](https://www.youtube.com/watch?v=dvDGkP3V9Z8&t=188s) **Preserving Human Skills Amid AI** - A speaker warns that overreliance on language models can erode essential human abilities, urging the maintenance of those skills to navigate uncertainty. ## Full Transcript
Okay. So the the first thing to say up
front is that there is a lot of
potential in AI AI and probably more
potential in AI than just in LLMs and
people tend to confuse the two. You know
large language models are not the
totality of what you can do with AI.
Um and the utility is high. I mean this
morning I was taking a transcript of a
presentation I gave. I was taking some
slides. I put that into an LLM and it
produced a really good summary based on
my original content which I can send to
the client. So that ability to to
summarize to synthesize is something
LLMs are good at because they use
improbability models based on training
data sets. The real problem for human
beings is that our own sense making
capability could effectively be replaced
by that. not by it doing what we do
well, but it by dumbing ourselves down
to the point where we actually don't see
the capability of humans. So the big
thing we're working on with the new
research program is to get the balance
between AI and and human based reasoning
decision- making and to understand not
only how that happens but the things
that human beings have to do over an
extended period in order to develop
judgment type capability.
The worry is the pace of change on LM's
their use for human tasks could actually
mean that we lose that capability or it
gets lost substantially certainly from
major companies and that's an issue
hence our development of audit tools.
Yeah. Um
AI LLMs in particular taken unfettered
and take for primary data collection are
always going to reduce our imagination.
They're going to reduce the
possibilities. They think inductively we
think abductively. We use metaphor. We
use art. We use imagination. We actually
use chemicals. We use senses and data we
don't yet fully understand to make
decisions under conditions of extreme
uncertainty.
Yeah. Where the past is not repeating in
in a way which can be predicted by a
framework or a model. So there are
things and rules that we need to create
around this. I blogged on this recently.
Things like stop talking about AI as if
it was a human being. Stop
anthropomorphizing it. Yeah. actually
make sure we practice things
independently of LLMs.
And certainly my own rule is not to use
it for primary data collection or
primary search because then I lose the
capacity to find things I didn't expect
to find. Remember looking through card
indexes in the university library and
finding a rather obscure essay by a guy
called Joseph Stalin when he was a
doctoral student at a college in Georgia
on toolsto what is art. Now that would
have never come up with an LLM because
it was an outlier. it wasn't a dominant
pattern. So we need to maintain search
capabilities. We need to maintain
production capabilities, synthesis and
representation. I can see AI doing a lot
more of that. But again, we need to
maintain our skill base. And probably
the best metaphor I can think of is I
make sure I climb a lot. I walk a lot.
I've always got GPS. I've got two GPS
systems. But about once a month, I make
sure I walk with a map and compass.
Otherwise, I will lose that skill. So
start to think about the skills that
human beings have they need to maintain
which while LLM might help with that,
augment it or act as a hybrid, we have
to maintain those skills for conditions
of uncertainty.
[music]