Good - Bad AI? c/o FT Magazine et al.

My source: Tim Harford, How can we tell good AI from bad? FT Magazine. 21 March 2026. #1168: pp.9-10.

    'Two new working papers address the tricky issue of verifying quality. In "Some Simple Economics of AGI", Christian Catalini, xiang Hui and Jane Wu (assisted, sometimes gratingly, by generative AI) propose the inevitable 2x2 matrix in which economic activity can be easy to automate, easy to verify, both or neither. Automatable, verifiable output is the stuff that computers do for us. The non-automatable stuff remains reassuringly artisanal.
    The difficult quadrant is where tasks seem easy to complete but are hard to check. Catalini, Hui and Wu call this the "runaway risk zone". It is not a reassuring label and it is not meant to be.
...
    In the second paper, "A model of Artificial Jagged Intelligence", Joshua Gans offers an analogy in which asking AI to perform task is like trying to cross a river over a network of planks supported by occasional pylons. The jagged frontier is represented by the fact that some planks are long and wobbly, while others are short and sturdy. Problem one: even if the planks are typically sturdy, the wobbly planks will require most of your time and attention. Problem two: if you can't predict in advance which planks will let you down, you may quite sensibly prefer to eschew the AI entirely and row yourself across the old-fashioned way.'

Catalini, C., Hui, X., & Wu, J. (2026). Some Simple Economics of AGI. arXiv preprint arXiv:2602.20946.

Joshua S. Gans, 2026. "A Model of Artificial Jagged Intelligence," NBER Working Papers 34712, National Bureau of Economic Research, Inc.

Previously: 'safety' : 'ai' : 'domain' : 'matrix'

AI - World Models

'What Is a World Model?

World models are neural networks that understand the dynamics of the real world, including physics and spatial properties. They can use input data, including text, image, video, and movement, to generate videos that simulate realistic physical environments. Physical AI developers use world models to generate custom synthetic data or downstream AI models for training robots and autonomous vehicles.'

https://www.nvidia.com/en-us/glossary/world-models/

In healthcare a 'world model' is slightly more expansive, hence the importance of experienced humans. We call these - nurses, doctors, physiotherapists, occupational therapists, social workers and many other professions, support workers and disciplines. They learn and train for many years and must continue to learn and unlearn throughout their careers. Their work and engagement is shaped and directed by human values, which are in turn informed by social change, evidence-based research, professional guidance, policy and law.

individual

|

INTERPERSONAL : SCIENCES

humanistic -------------------------------------------  mechanistic

SOCIOLOGY : POLITICAL

|

group-population

	WORLD MODEL

Modelling turbulence in flight: and other situations?

'Thomas Q. Carney, a retired professor of aviation technology at Purdue University in Indiana who has logged more than 11,000 flight hours as a pilot, said, "The better the model, the more it captures of the particular turbulent field, then the better the forecast, which is what the pilot is going to use."'

individual

|

INTERPERSONAL : SCIENCES

humanistic -------------------------------------------  mechanistic

SOCIOLOGY : POLITICAL

|

group-population

complexity and turbulence here ...	Birnir, B., & Angheluta, L. (2025). Scaling of Lagrangian structure functions. Phys. Rev. Res., 7(2), 023225.  https://doi.org/10.1103/PhysRevResearch.7.023225
here	and here ...

'To try to make sense of chaos, Dr. Birnir worked with Luiza Angheluta-Bauer, a theoretical physicist at the University of Oslo, to devise a model that combines two different methods for observing turbulence: what are known as Lagrangian and Eulerian mechanics. Experts say that neither framework can fully explain how turbulence works. 

That's because these two frameworks look at fundamentally different aspects of a turbulent system. In Lagrangian mechanics. researchers observe a simple particle. while in the Eulerian framework they look at a single point in space. Put simply, Lagrangian mechanics is like watching a leaf flow downriver, subject to the whims of eddies in the water. On the other hand, Eulerian mechanics is like watching a rock that protrudes from the river's surface and studying how the turbulence of the water moves around that fixed point.

Lagrangian turbulence is trickier to model because it requires an understanding of how a lone particle will behave. That lone particle will "execute the most complicated motion that you could imagine," Dr. Birnir said.

Knowing how each type of turbulence fits into the bigger picture is akin to selecting the appropriate lens for a microscope, since both are highly dependent on perspective. "Same turbulence, different stories," said Tomek Jaroslawski a postdoctoral fellow at the Center of Turbulence Research at Stanford University.' Nazaryan, p.12.

Nazaryan, A. A new theory of in-flight turbulance lands, Science: The New York Times International Edition, October 15, 2025, p.12.

[Subscription: https://www.nytimes.com/2025/09/24/science/physics-airplanes-turbulence.html ].

See previously: 'care architecture' : 'Serres' : 'complexity' - and, if needed 'spiritual'

c/o Ridgway - assimilation & accommodation

An approach to testing and teaching

'Consider the problems of acquiring new knowledge; two ideas are particularly important. The first is that existing knowledge and conceptual structures affect the way that new materials are perceived and learned; the second is that new experiences bring about changes in our knowledge and conceptual structures. These two processes are referred to as assimilation (analogous to the way in which the stomach digests food, whose later structure cannot be recognized as being similar to its earlier structure) and accommodation (analogous to the way that the pupil of the eye adjusts itself different light levels). Biological analogies make it easier to understand these concepts but leave unresolved the question of when one accommodates and when one assimilates. In general, one accommodates (i.e. changes one's conceptual structures) when fresh knowledge provides strong challenges to what is in mind; so dramatic counter-examples to currently held beliefs, or coherent patterns in the world, which cannot be explained by existing beliefs are both likely to bring about accommodation. Assimilation (interpreting new events in terms of old ideas) can be made to work when the number of counter-examples to predictions made from old beliefs are rather low. It follows, therefore, that if misconceptions are to be remedied (i.e. the mechanism of assimilation is overcome and the mechanism of accommodation stimulated) a representative sample of questions in the domain of interest is unlikely to have the desired effect because the number of cases which violate pupils' misconceptions and which therefore  might cause accommodation, will be relatively small. To foster accommodation it is necessary to provide dramatic examples which violate current conceptions and to provide these examples in quantity. Examples which are most likely to be dramatic are those in which it is obvious to pupils that the results they are obtaining using particular misconceptions are at variance with what they 'know to be true' from everyday experience.' pp.46-47.

Jim Ridgway (1988). Assessing Mathematical Attainment. Chapter 3, Using Test Results.Windsor Berkshire. NFER-NELSON. pp.40-52. [Ack. length of quote - See also: https://www.nfer.ac.uk/ ]

Thanks to Lancaster Univ. Library.

Previously 'assimilation' : 'accommodation' : 'math'

HC@AIxIA AI&Health May Seminar: 'Building Trustworthy AI for Health'

Dear Madam/Sir,

This is to officially announce the MAY 2025 seminar of the "AI & Health: Seminars 2025" series as hosted by HC@AIxIA, i.e., the "Artificial Intelligence for Healthcare" working group of the Italian Association for Artificial Intelligence.

*** Save the date: 12 MAY 2025. 3:30 CET ***

We hope you will attend and participate in the discussion on the relevant topics that will be presented and by our speakers. Feel free to share this with those potentially interested.

Link for participating: https://bit.ly/hc-2025-05 (PLEASE CHECK the site https://aixia.it/en/gruppi/hc/ for any changes or updates) - Please find some details below, and a poster attached.

== May 2025 seminar ==
2025 May 12 - 3:30PM CET

Prof. Barbara Di Camillo, Department of Information Engineering, University of Padova, Italy

Title: Building Trustworthy AI for Health: Robust and Generalizable Models with a Focus on Challenging Cases

Abstract: To make artificial intelligence useful in real clinical settings, it is crucial that AI tools follow the principles of trustworthy AI. This means ensuring the accuracy and reliability of the results, defining the areas where the results are valid, and making sure the predictions are understandable. This approach ensures that humans remain at the center of the decision-making process. BRAINTEASER (https://brainteaser.health/) is a data science project that uses artificial intelligence to help patients with amyotrophic lateral sclerosis (ALS) and multiple sclerosis (MS), along with their doctors. In the presentation, I will show how, during the development of the machine learning methods within the project, we took into account the robustness and generalizability of the model, and how we identified and characterized the subjects for whom making predictions is more challenging.

Short Bio: Barbara Di Camillo is full professor in computer science with the Department of Information Engineering, University of Padova. Her research activity is centered in the development and application of advanced modeling, data mining and machine learning methods for high-throughput biological data analysis in the field of Bioinformatics and Systems Biology. In particular, she has developed and applied different methods for robust biomarker discovery, predictive modeling and clustering of clinical data and next generation sequencing data. She has also a great expertise in the development and application of differential equation based models, Boolean and Bayesian Networks for modeling the relationships between the variables and the pathways along which they influence the disease progression.

Flyer: https://drive.google.com/open?id=15badJ0Nfj8lrA0gZdFc9aOGeR3pelms-&usp=drive_fs

====

Some notes
Serving as coordinators of the working group on AI for Healthcare of the Italian Association of Artificial Intelligence (AIxIA, see: https://aixia.it/en/gruppi/hc/), part of our commitment consists of fostering contamination and collaboration between AI researchers and experts and operators in Medicine and Healthcare; in particular, we aim to contribute in building a two-way road for informing healthcare operators about AI results and opportunities, while also raising awareness among AI researchers about challenges and problems in medicine and healthcare.
Therefore, the 2025 seminar series, in the trail of the 2024 edition, will feature a number of experts presenting research results, projects, best practices, ideas, and more to a mixed audience of AI researchers and healthcare operators.

Thank you for your interest in the AI & Health seminar series and the HC@AIxIA working group, and see you soon!

Sincerely,
Francesco Calimeri, Mauro Dragoni, Fabio Stella
(coordinators of the HC@AIxIA working group)

Harnessing the Power of Artificial Intelligence to Improve Outcomes for Patients with for Long-Term Health Conditions

 Dear Colleagues and Friends,

We are organising a Special Session: Harnessing the Power of Artificial Intelligence to Improve Outcomes for Patients with for Long-Term Health Conditions

(https://aiih.cc/lthc/) in the International Conference on AI in Healthcare (AIiH), 8-10 September 2025, Jesus College, University of Cambridge.

We would like to accept both full length papers (12 pages plus references) and short abstracts (up to 5 pages including references) for special sessions. Submission guideline can be found here, including paper templates in both Word and LaTeX: https://aiih.cc/paper-submission/

The accepted full papers and abstracts will be published in the Springer LNCS volumes.

Full Paper submission deadline:            Friday 11 April 2025

Abstract submission deadline:               Monday 30 June 2024

We are looking forward to meeting you.

Best wishes

Shang-Ming Zhou

Professor in e-Health | Faculty of Health | University of Plymouth | PL4 8AA | UK.

Email :  shangming.zhou AT plymouth.ac.uk; smzhou AT ieee.org

https://www.plymouth.ac.uk/staff/shang-ming-zhou

https://www.plymouth.ac.uk/research/centre-for-health-technology

History: Hepatoscopy and the Census

INDIVIDUAL

|

     INTERPERSONAL    :     SCIENCES               

HUMANISTIC --------------------------------------  MECHANISTIC      

SOCIOLOGY  :   POLITICAL 

|

GROUP

"Perhaps it was merely an oversight of this historian that he failed to mention another religious practice, hepatoscopy. For this, the Romans down to imperial times relied on Etruscans. Divination from the liver of sacrificed animals was a most complicated business.

To get the interpretation right required a sound knowledge not only of astronomy but also of anatomy and pathology. In 1877 at Piacenza a bronze model of a liver was found. It was evidently a model used in a priests' school for teaching the art of haruspicy, and of a fairly late period. On the upper surface are marked sixteen divisions in which are inscribed the names of some thirty gods, some more than once. Each of the sixteen divisions corresponded to a section of the heavens, for the Etruscans regarded the liver as an image of the cosmos, as a microcosm of the universe. The liver, it was believed, exactly reproduced the heavenly firmament with its sacred laws and controlling divinities and its division, first into four, by the intersecting axes and then into sixteen subdivisions." p.87.

The Etruscans

"Censuses for purposes of taxation and military service still exist. But not many know that it was an Etruscan king who first legally established such a registration in Rome, from whence, with the rise of the Roman empire, it spread throughout Europe."  "What this measure was is stated by Livy thus: 'The population was divided into classes and 'centuries' according to a scale based on the census, and suitable for both peace and war." p.146.

Keller, W. (1975) The Etruscans. Norfolk: Book Club Associates. [Cover image: AbeBooks]

AI Anxiety and the Public Letter in Hodges' model

INDIVIDUAL
|
    INTERPERSONAL  :   SCIENCES              
HUMANISTIC --------------------------------------  MECHANISTIC
SOCIOLOGY :   POLITICAL
|
GROUP

'doomerism' fear, belief, rationality artificial intelligence AI anxiety reflection, critical thinking	echo extinction biological intelligence pace of change past Speed NOW Time future
public public understanding of science 'public letter' collective fear - history life: family, work, leisure ..	policy governance motivation: public good, marketing, business, profit, power, control . . .

My source:

"Something strange has appeared in our sky. Let’s get a sense of what this alien is before we blow it out of the sky. Maybe it’s beautiful!"
https://t.co/UKwZCkjYrN

No apocalypse for now! An excellent, balanced piece on AI anxiety from @StephenMarche @guardian

Let's think more… pic.twitter.com/iJvPlzzK1q

— AI (@DeepLearn007) May 15, 2023

ERCIM News No. 129 Special Theme: "Fighting Cybercrime"

 Dear ERCIM News reader,

Fighting Cybercrime is the special theme of ERCIM News No.129, just published at https://ercim-news.ercim.eu/

This issue's Special Theme contains many promising concepts, methods and solutions that contribute to the defense against and fight against cybercrime. This special theme has been coordinated by our guest editors Florian Skopik (AIT Austrian Institute of Technology) and Kyriakos Stefanidis  (ISI).

Thank you for your interest in ERCIM News. Feel free to forward this message to anyone who might be interested. We are also happy if you follow us and talk about us on twitter @ercim_news and other social media.

Next issue:
No. 130,  July 2022
Special Theme: "Assistive and Inclusive Technologies" (submissions welcome!)

---

ERCIM News is published quarterly by ERCIM, the European Research Consortium for Informatics and Mathematics. With the printed and online edition, ERCIM News reaches more than 10000 readers.

About ERCIM

ERCIM - the European Research Consortium for Informatics and Mathematics - aims to foster collaborative work within the European research community and to increase co-operation with European industry. Leading European research institutes are members of ERCIM. ERCIM is the European host of W3C.

https://www.ercim.eu/
https://twitter.com/ercim_news

Peter Kunz                      	
ERCIM Office
2004, Route des Lucioles
BP93
F-06902 Sophia Antipolis Cedex 

Mechanisms, Individuals and Collective

Last week I joined a webinar:

 

https://www.biokoinos.org/events

BioKoinos Research Seminar

Saúl Pérez González & Elena Rocca - 

Evidence-based and patient-centred clinical practice. The bridging role of evidence of mechanism.

  

I have mapped some of the concepts to Hodges' model. A link to the paper is included below. The session was very helpful, especially with ongoing developments in genomics and personalised medicine.

 

INDIVIDUAL

|

 INTERPERSONAL    :     SCIENCES               

HUMANISTIC --------------------------------------  MECHANISTIC      

SOCIOLOGY  :   POLITICAL 

|

GROUP

patient-centred evidence safety of a certain intervention for a particular patient safety efficacy	individual - epidemiology - population patient-centred evidence safety of a certain intervention for a particular patient biology pathophysiological mechanism prediction ... confirming and disconfirming - depending on whether or not it supports that certain epidemiological results apply to the single patient.
evidence safety efficacy	evidence safety efficacy

Pérez-González, S., & Rocca, E. (2022). Evidence of Biological Mechanisms and Health Predictions: An Insight into Clinical Reasoning. Perspectives in Biology and Medicine 65(1), 89-105.  

http://philsci-archive.pitt.edu/20364/1/Postprint%20Evidence%20of%20biological%20mechanisms%20and%20health%20predictions.pdf

 

Occam's razor (i)

"2. The virus of encyclopedism in definitions should be
recognized and resisted. There is a role and a place for
illustrative detail, but that role and place is not in a
definition. Social science is content to tolerate excep-
tions to its theories—it is satisfied with explanations
covering most cases, not each and every one. In the
definition of strategic concepts, more detail inevitably
promotes less clarity and therefore less understanding.
William of Occam should be regarded as the patron
saint of wordsmithing for strategic conceptualization.

3. Ideas matter: Concepts for theory have practical
consequences. The way in which we behave strategi-
cally is not dictated strictly by the way in which we
conceptualize its challenges and intellectually order
our possible responses, but our concepts educate our
perception and interpretation of events, and they find
expression in the doctrine that shapes our behavior.52
Of course, strategic behavior should be adaptable to
unanticipated events, but frequently it is not. Strategic
and military culture can and does change, but at any
one time it is going to help mold action now in ways
organized doctrinally in the light shed by authorita-
tive strategy concepts." (p.46).

 
Gray, C. S. (2012). CATEGORICAL CONFUSION?: THE STRATEGIC IMPLICATIONS OF RECOGNIZING CHALLENGES EITHER AS IRREGULAR OR TRADITIONAL. Strategic Studies Institute, US Army War College. http://www.jstor.org/stable/resrep11257

Of Bodies and Persons

individual
|

INTERPERSONAL : SCIENCES

humanistic ----------------------------------------------- mechanistic

SOCIOLOGY : POLITICAL

|

group

	Three-body problem
Three-Person problem

 

https://www.britannica.com/science/three-body-problem

 

Happy New Year: Nostradamus c/o Al Stewart

"A Man for All Seasons" ...
"A Woman ... and other gender identities ..."
"A Child ..."
"A model ..."

"A model for Past, Present, Future"

A series of Frames - Janus-like view:

Pre-morbid, Diagnosis, Prognosis ...

Clinical and Case -
Assessment
Planning
Formulation
Evaluation
Engagement

individual

|

INTERPERSONAL : SCIENCES

humanistic ----------------------------------------------- mechanistic

SOCIOLOGY : POLITICAL

|

group - population

Nostradamus - Al Stewart
From the Album: Past, Present, Future

Steve Coleman And Five Elements, Functional Arrhythmias

individual

|

INTERPERSONAL : SCIENCES

humanistic --------------------------------------- mechanistic

SOCIOLOGY : POLITICAL

|

group

20/20 vision minus 1, 2, or 4 blind spots....

Basic science classes very quickly introduce students to 'the eye' and vision.

This includes the simple experiment that can be done which reveals the blind spot.

Here's an example from (what was) Service Works Consulting:

Consulting companies pride themselves on knowledge, foresight and expertise. This is fine if there is only one blind spot to deal with, but of course we each have two. From the perspective of project management there are many more blind spots.

Hodges' model suggests there are potentially four;
five even if we include the spiritual domain.

A benefit of using Hodges' model* is that just as in normal vision our two eyes overlap and compensate for the blind spots; so attention paid to just one additional care domain may diminish the impact of care domain blindness.


3 out of 4 domains - is better (more holistic) than - 2 out of 4.

The problem in health and social care, is that any one of four blind spots (ironically the site of the optic nerve bundle) can become a fuse for trouble or disaster.

If you have the gift of vision - best to use it.^#
*cognitively or deliberately on paper or computer

Image source with many thanks: John Eric Hughes

Additional links:
Visual acuity: http://en.wikipedia.org/wiki/Visual_acuity

^#Wish I had and could do that!