Researcher applying cognitive science to the intersection of people, technology, and work | Ph.D. Human Factors
Joined November 2019
- Tweets51
- Following99
- Followers14
- Likes167
Honored to have recently spoken about my research supporting cognitive performance in Intelligence Analysis.
I discuss how Signal Detection Theory provides an important perspective on cognitive work in domains that are characterized by high uncertainty and complexity
New podcast episode out now! 🎙
Meet the 2022 & 2024 Gary A. Klein and Robert R. Hoffman Award winners and hear about the work behind their achievements. Plus, some lighthearted Q&A!
Listen here:
naturalisticdecisionmaking.o…
Jordan Haggit retweeted
On the ubiquity of the #cognitivebias bias.
linkedin.com/posts/brianmoon…
Jordan Haggit retweeted
Want to learn how AI planning, machine learning, and interaction design can help people make better decisions?
Register for #NDM2024 to hear @tmiller_uq, Professor of Artificial Intelligence @UQSchoolEECS, discuss these hot topics & more!
So much wrong with system 1/system 2 when applied to complexities of the real world.. reductionist at best when considering factors like expertise, time constraints, amount of potentially relevant info, etc, etc
behavioral economists will keep using dual process theory/systems 1 and 2 because IT'S SO USEFUL, even if it's not real - we can't stop them but we can be aware that it's fake/meaningless scholar.google.com/scholar?c…
Jordan Haggit retweeted
IFKYK. @NDMAssn coffee chats are attracting quite the crowd of heavy hitters (@DrFadde joke!).
Join today & get in here!
@JBorders & Gary Klein from @ShadowBoxTrain
Robert Hoffman from @TheIHMC
@l_militello @MCarrawayPrehD
@toomasroolaid @JephsonWendy @HaggitJordan
Assuming this is heavily influenced by kahneman and the bias literature, which is based on a misplaced framing of human abilities @farnamstreet highly recommend @teppofelin paper on “surprise hacking”. Or Gary Klein’s work on expert intuition
The differences are mitigated from experience hacking and modularity of furniture. Suggesting novelty comes from exploring the right hypotheses/frames that allow new affordances to arise. A combo of schema, tool, environment. @teppofelin
clever Management Science paper on how customers hack, repurpose and modify IKEA products. Note the importance of "problem-first search" in their findings.
Chan & Lim: Exaptations in IKEA Hacks doi.org/10.1287/mnsc.2022.44… @IKEA @INFORMS
3/3 How we prioritize and present the information is also critical to consider. People should understand the capabilities and limitations of AI tools and we should present information in ways that makes those capabilities observable and transparent.
2/3 Although the context will vary, we have a good understanding of what data are important to address these challenges
AI tools should be used to extend the ability of people to navigate complex data or make critical patterns apparent that were previously unseen.
1/3 How can we use AI tools to enhance human work and performance?
In high stakes and safety critical domains we can look to cognitive systems engineering for guidance.
Some key challenges people encounter include: planning, monitoring, predicting, adapting, and learning.
Jordan Haggit retweeted
Looking for more information on Naturalistic Decision Making?
The NDM community has generated various tools you can explore, including methods, models, guidance, and research support.
Explore here!👇
naturalisticdecisionmaking.o…
Jordan Haggit retweeted
Explicit knowledge is just the tip of the iceberg when it comes to understanding #expertise.
Tacit knowledge involves pattern recognition, perceptual discriminations, mental models, and judgments of typicality.
An incredible example of Ashby’s law of requisite variety that is a great reminder that we can only meet and support work in complex systems with complexity. youtu.be/mDRcLAkRZ50
Jordan Haggit retweeted
Mile Two Cognitive Systems Engineering Manager Layla Akilan gave her TEDxDayton talk on user-centered design last October.
It's finally up on YouTube! Enjoy: buff.ly/3Jx5ZJR
The narrative of “Full swing” on Netflix is PGA (golf legacy) vs LIV (family legacy). Netflix could have benefited from some better interview techniques to unpack that false dichotomy. That probably is less entertaining for most to be fair
Incentive structure is myopic. Developers/engineers are concerned with winning the feature race. Doyle's catch (Alderson & Doyle, 2010) highlights the risks of demos being misleading and masking their brittleness to real world complexity. Need to design for resilience
We need transparency at all levels for autonomous systems. Anthropomorphism has a high risk of confusing users about its capabilities and limitations. At best It’s misleading and at worst it’s consequences are legitimately criminal.
People mistaking AI systems for people is about to be a huge problem. Going out of your way to make your AI system pretend to be an appealing person is irresponsible.
SusanC, commenting at my site: betterwithout.ai/scary-langu…
Jordan Haggit retweeted
John Flach, PhD giving his talk at the @NDIAToday Human Systems Conference today!
#ndia #humansystems
ALT John at the podium giving his talk