An Analytical Framework for Multi-Theoretic Ethical Stress Test (MTEST): Ethical Analytics on Sovereign AI and Artificial General Intelligence

Javed I.Khan, Sharmila Rahman Prithula
Proceedings of the The 39th Canadian Conference on Artificial Intelligence, PMLR 318:127-138, 2026.

Abstract

The rise of pervasive computing and the pursuit of Artificial General Intelligence (AGI) have moved AI ethics from philosophical debate to a core requirement for global governance. However, ethical evaluation remains a highly subjective task, largely inaccessible to general technologists, and often ad-hoc- due in part to the absence of any structured, pluralistic framework capable of assessing alignment across diverse moral perspectives. This paper presents MTEST11- a Multi-Theoretic Ethical Stress Test that offers a systematic and quantifiable approach to evaluating the ethical soundness of propositions by alignment checks against most influential ethical theories (THE11), including utilitarianism, deontology, rights-based ethics, Rawlsian justice, virtue ethics, and others. The framework, while intentionally simplified for functional application, offers sufficient structure to support systematic quantitative analysis. It measures (i) ethical alignment of propositions, (ii) cross-theoretic consensus on propositions, (iii) moral congruence of individual theories on a proposition set, iv) and shields against any ethical blind spots of any single framework. It also reveals the (v) ethical value anchor set- the set of universally recognized ethical values on which a proposition is supported or contradicted. We demonstrate the utility of MTEST11 by applying it to perform quantitative and qualitative analysis of 14 provocative policy propositions from various sides of ongoing global debate on artificial general intelligence (AGI).

Cite this Paper

BibTeX
@InProceedings{pmlr-v318-i-khan26a, title = {An Analytical Framework for Multi-Theoretic Ethical Stress Test (MTEST): Ethical Analytics on Sovereign AI and Artificial General Intelligence}, author = {I.Khan, Javed and Prithula, Sharmila Rahman}, booktitle = {Proceedings of the The 39th Canadian Conference on Artificial Intelligence}, pages = {127--138}, year = {2026}, editor = {Bouzar-Benlabiod, Lydia and Leung, Carson}, volume = {318}, series = {Proceedings of Machine Learning Research}, month = {25--29 May}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v318/main/assets/i-khan26a/i-khan26a.pdf}, url = {https://proceedings.mlr.press/v318/i-khan26a.html}, abstract = {The rise of pervasive computing and the pursuit of Artificial General Intelligence (AGI) have moved AI ethics from philosophical debate to a core requirement for global governance. However, ethical evaluation remains a highly subjective task, largely inaccessible to general technologists, and often ad-hoc- due in part to the absence of any structured, pluralistic framework capable of assessing alignment across diverse moral perspectives. This paper presents MTEST11- a Multi-Theoretic Ethical Stress Test that offers a systematic and quantifiable approach to evaluating the ethical soundness of propositions by alignment checks against most influential ethical theories (THE11), including utilitarianism, deontology, rights-based ethics, Rawlsian justice, virtue ethics, and others. The framework, while intentionally simplified for functional application, offers sufficient structure to support systematic quantitative analysis. It measures (i) ethical alignment of propositions, (ii) cross-theoretic consensus on propositions, (iii) moral congruence of individual theories on a proposition set, iv) and shields against any ethical blind spots of any single framework. It also reveals the (v) ethical value anchor set- the set of universally recognized ethical values on which a proposition is supported or contradicted. We demonstrate the utility of MTEST11 by applying it to perform quantitative and qualitative analysis of 14 provocative policy propositions from various sides of ongoing global debate on artificial general intelligence (AGI).} }
Endnote
%0 Conference Paper %T An Analytical Framework for Multi-Theoretic Ethical Stress Test (MTEST): Ethical Analytics on Sovereign AI and Artificial General Intelligence %A Javed I.Khan %A Sharmila Rahman Prithula %B Proceedings of the The 39th Canadian Conference on Artificial Intelligence %C Proceedings of Machine Learning Research %D 2026 %E Lydia Bouzar-Benlabiod %E Carson Leung %F pmlr-v318-i-khan26a %I PMLR %P 127--138 %U https://proceedings.mlr.press/v318/i-khan26a.html %V 318 %X The rise of pervasive computing and the pursuit of Artificial General Intelligence (AGI) have moved AI ethics from philosophical debate to a core requirement for global governance. However, ethical evaluation remains a highly subjective task, largely inaccessible to general technologists, and often ad-hoc- due in part to the absence of any structured, pluralistic framework capable of assessing alignment across diverse moral perspectives. This paper presents MTEST11- a Multi-Theoretic Ethical Stress Test that offers a systematic and quantifiable approach to evaluating the ethical soundness of propositions by alignment checks against most influential ethical theories (THE11), including utilitarianism, deontology, rights-based ethics, Rawlsian justice, virtue ethics, and others. The framework, while intentionally simplified for functional application, offers sufficient structure to support systematic quantitative analysis. It measures (i) ethical alignment of propositions, (ii) cross-theoretic consensus on propositions, (iii) moral congruence of individual theories on a proposition set, iv) and shields against any ethical blind spots of any single framework. It also reveals the (v) ethical value anchor set- the set of universally recognized ethical values on which a proposition is supported or contradicted. We demonstrate the utility of MTEST11 by applying it to perform quantitative and qualitative analysis of 14 provocative policy propositions from various sides of ongoing global debate on artificial general intelligence (AGI).
APA
I.Khan, J. & Prithula, S.R.. (2026). An Analytical Framework for Multi-Theoretic Ethical Stress Test (MTEST): Ethical Analytics on Sovereign AI and Artificial General Intelligence. Proceedings of the The 39th Canadian Conference on Artificial Intelligence, in Proceedings of Machine Learning Research 318:127-138 Available from https://proceedings.mlr.press/v318/i-khan26a.html.

Related Material