About

Discernment, answerable to evidence.

I'm Kareem Soliman. My background is in psychological science and organisational psychology, with a Graduate Diploma in Organisational Behaviour and a Bachelor of Business (Finance). I build with AI every day, and I measure what it does.

01

Consequential measurement

Lately my work keeps returning to one question: when a measurement changes how people are treated, how do we make it fair?

It runs through everything in the Lab: how blame is apportioned after a harm, how drivers are reported, how AI judges are trusted, how a filing becomes a price. Each is a reading that someone will act on.

02

Why firāsa

فِراسة

firāsa

Firāsa is an Arabic word associated with discernment: reading inward qualities through outward signs. Its history includes intuitive and religious insight, learned interpretation and physiognomy. These traditions differ; their existence does not establish the truth of their claims.

For me, the name connects the question that first drew me towards psychology to the discipline I later found in measurement: how much can we responsibly infer about something we cannot observe directly? An action is visible; another person’s reasons are not. A pattern appears in data; its cause does not arrive beside it. Code behaves in one way; the quality we care about is still an interpretation of many signs.

Firāsa does not mean privileged access to the truth. My contemporary reading treats it as disciplined discernment from signs: an effort to notice what a surface might disclose while remaining answerable for the inference. Modern data analytics expands the signs we can examine and the patterns we can compare. It does not decide what those patterns mean. Empathy opens possibilities. Theory helps organise them. Measurement asks which reading the evidence can support, which alternatives remain plausible and what would change my mind.

Firāsa without measurement can become projection. Measurement without firāsa can become counting. The first fails by mistaking a compelling interpretation for evidence; the second fails by recording what is easy to count without asking whether it bears on the thing we need to understand. I want the work to keep both failures visible.

There is no fixed ratio between discernment and measurement. The appropriate burden depends on the decision: what it will ask of people, how consequential an error would be, how quickly a judgement must be made and whether it can later be revised. A tentative question may justify a light instrument and an open reading. A decision about safety, opportunity or another person’s future demands stronger evidence, clearer alternatives and a more exact account of uncertainty.

That is why the word has become increasingly important to me. It holds together two instincts I do not want to separate: the desire to understand what lies behind an observable sign, and the obligation to discover when my most compelling interpretation is wrong. A good reading should remain answerable to the person, the data and the next observation. This is my interpretation of the name, not a claim that its varied history already contained modern measurement science.

Sources for the word and its history: Lane’s Arabic-English Lexicon ↗ · Ghersetti, Firāsa and Intelligence (1996) ↗

03

The discipline

I ask what would justify relying on a reading. That means distinguishing an observation from an interpretation, testing the instrument and choosing a level of rigour proportionate to the consequences. The aim is calibration, not a universal fifty-fifty compromise between judgement and data.

Measurement should earn its place. A more elaborate analysis is useful only if it improves understanding or decisions enough to justify its burden.

04

One research outlook

The research runs in four strands: judging people fairly, measuring AI honestly, reading systems and money, and building open tools others can use. Each approaches a different part of the same problem: learning about what is not directly observable.

Organisational psychology connects the work to the design and evaluation of human and AI teams, and to the practical question of how a team actually takes up a new tool.

05

About me

I am an independent researcher in Sydney with a background in psychological science and organisational behaviour. My work connects measurement, AI evaluation and decision-making under uncertainty.

I use AI throughout Firasa’s research, development and writing. I remain responsible for the questions, the claims and the decisions.

BPsychSc (Hons), ACAP · Graduate Diploma in Organizational Behaviour, Macquarie · BBus, UTS

Explore the Lab ↗