{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,21]],"date-time":"2026-04-21T04:48:11Z","timestamp":1776746891763,"version":"3.51.2"},"reference-count":32,"publisher":"World Scientific Pub Co Pte Ltd","issue":"01","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Int. J. Semantic Computing"],"published-print":{"date-parts":[[2026,3]]},"abstract":"<jats:p>Safety-critical sectors such as power transmission and autonomous mobility are not replacing their deterministic controllers; they are adding artificial intelligence (AI) layers to existing systems to gain additional, sustainable business value. This approach increases the socio-technical complexity, creating pressure for a structured means to compare benefits, oversight effort and risk over time. In this descriptive study, we introduce Epistemic Gain [Formula: see text], a scalar derived from the [Formula: see text] model that links foresight gains to human oversight and system friction. Within this framework, [Formula: see text] is treated as a necessary condition for epistemically sustainable scaling. We then formulate a conjectured governance-level Law of Diminishing Returns that holds up to a Scaling Failure Threshold, beyond which marginal upgrades begin to destroy value. Drawing on recent empirical studies, we further sketch the [Formula: see text] model and show how [Formula: see text] can be displayed in software development lifecycle dashboards. This paper extends the earlier IEEE [Formula: see text] 2025 conference paper in three main directions: (i) provide the theoretical foundation of the M-Vector and formalize it as the explicit epistemic state [Formula: see text] underpinning the [Formula: see text] model; (ii) introducing semantic instability [Formula: see text] and epistemic drift [Formula: see text] as properties inspired by causal representation learning, used here as an AI-safety and governance lens; and (iii) identifying canonical regions of [Formula: see text]-space for deterministic, vital, symbolic, sub-symbolic and generative AI systems. The overarching aim is to formalize the theoretical basis of [Formula: see text] and its time-variant machine form, enabling it to serve as a governance indicator for when to scale, optimize, or pause AI deployments.<\/jats:p>","DOI":"10.1142\/s1793351x26410011","type":"journal-article","created":{"date-parts":[[2026,2,13]],"date-time":"2026-02-13T09:55:41Z","timestamp":1770976541000},"page":"5-30","source":"Crossref","is-referenced-by-count":0,"title":["Epistemic Gain in M\u2013Space: A Metric for AI Governance in Complex Systems"],"prefix":"10.1142","volume":"20","author":[{"ORCID":"https:\/\/orcid.org\/0009-0006-0294-6264","authenticated-orcid":false,"given":"Generoso","family":"Immediato","sequence":"first","affiliation":[{"name":"Hitachi Rail, Naples, Italy"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"219","published-online":{"date-parts":[[2026,3,7]]},"reference":[{"key":"S1793351X26410011BIB001","doi-asserted-by":"publisher","DOI":"10.1080\/15236803.2006.12001438"},{"issue":"1","key":"S1793351X26410011BIB002","first-page":"2025","volume":"19","author":"Immediato G.","journal-title":"Intersect: Stanford J. Sci. Technol. Soc."},{"key":"S1793351X26410011BIB003","volume-title":"Proc. 2025 Artificial Intelligence for Business","author":"Immediato G.","year":"2025"},{"key":"S1793351X26410011BIB004","unstructured":"D. Amodei, C. Olah, J. Steinhardt, P. Christiano, J. Schulman and D. Man\u00e9, Concrete problems in AI safety, arXiv:1606.06565 (2016), https:\/\/arxiv.org\/abs\/1606.06565."},{"key":"S1793351X26410011BIB005","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2021.3058954"},{"key":"S1793351X26410011BIB006","unstructured":"Y. Akhauri, B. Lewandowski, C.H. Lin, A. N. Reyes, G. C. Forbes, A. Wongpanich, B. Yang, M. S. Abdelfattah, S. Perel and X. Song, Performance prediction for large systems via text-to-text regression, arXiv:2506.21718 (2025), https:\/\/arxiv.org\/abs\/2506.21718."},{"key":"S1793351X26410011BIB007","unstructured":"K.H. Huang, A. Prabhakar, O. Thorat, D. Agarwal, P. K. Choubey, Y. Mao, S. Savarese, C. Xiong and C.S. Wu, CRMArena-Pro: Holistic assessment of LLM agents across diverse business scenarios and interactions, arXiv:2505.18878 (2025), https:\/\/arxiv.org\/abs\/2505.18878."},{"key":"S1793351X26410011BIB008","doi-asserted-by":"crossref","unstructured":"P. Shojaee, I. Mirzadeh, K. Alizadeh, M. Horton, S. Bengio and M. Farajtabar, The illusion of thinking: Understanding the strengths and limitations of reasoning models via the lens of problem complexity, arXiv:2506.06941 (2025).","DOI":"10.70777\/si.v2i6.15919"},{"key":"S1793351X26410011BIB009","unstructured":"C. Opus and A. Lawsen, Comment on \u201cThe illusion of thinking: Understanding the strengths and limitations of reasoning models via the lens of problem complexity\u201d, arXiv:2506.09250 (2025), https:\/\/arxiv.org\/abs\/2506.09250."},{"key":"S1793351X26410011BIB010","unstructured":"S. Bubeck, C. Coester, R. Eldan, T. Gowers, Y. T. Lee, A. Lupsasca, M. Sawhney, R. Scherrer, M. Sellke, B. K. Spears, D. Unutmaz, K. Weil, S. Yin and N. Zhivotovskiy, Early science acceleration experiments with GPT-5, arXiv:2511.16072 (2025)."},{"key":"S1793351X26410011BIB011","doi-asserted-by":"crossref","unstructured":"G. Immediato, Epistemic trade-off: An analysis of the operational breakdown and ontological limits of \u201ccertainty-scope\u201d in AI, arXiv:2508.19304 (2025).","DOI":"10.36227\/techrxiv.175356175.57564616\/v3"},{"key":"S1793351X26410011BIB012","doi-asserted-by":"publisher","DOI":"10.1016\/0160-2896(95)90024-1"},{"key":"S1793351X26410011BIB013","doi-asserted-by":"publisher","DOI":"10.1613\/jair.1.11388"},{"key":"S1793351X26410011BIB014","unstructured":"CENELEC. EN 50126:2017 \u2014 Railway Applications \u2014 The Specification and Demonstration of Reliability, Availability, Maintainability and Safety (RAMS)."},{"key":"S1793351X26410011BIB015","unstructured":"International Electrotechnical Commission (IEC). IEC 61508 (Parts 1\u20137), Edition 2.0 (2010) \u2014 Functional safety of electrical\/electronic\/programmable electronic safety-related systems."},{"key":"S1793351X26410011BIB016","volume-title":"Cybernetics: Or Control and Communication in the Animal and the Machine","author":"Wiener N.","year":"1948"},{"key":"S1793351X26410011BIB017","doi-asserted-by":"publisher","DOI":"10.5962\/bhl.title.5851"},{"key":"S1793351X26410011BIB018","doi-asserted-by":"publisher","DOI":"10.1093\/analys\/58.1.7"},{"key":"S1793351X26410011BIB019","doi-asserted-by":"publisher","DOI":"10.7551\/mitpress\/1881.001.0001"},{"key":"S1793351X26410011BIB020","unstructured":"J. Kaplan, S. McCandlish, T. Henighan, T. B. Brown, B. Chess, R. Child, S. Gray, A. Radford, J. Wu and D. Amodei, Scaling laws for neural language models, arXiv:2001.08361 (2020)."},{"key":"S1793351X26410011BIB021","first-page":"30048","volume":"35","author":"Hoffmann J.","year":"2022","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"S1793351X26410011BIB022","doi-asserted-by":"publisher","DOI":"10.1145\/3531146.3533088"},{"key":"S1793351X26410011BIB023","unstructured":"L. Weidinger, M. Rauh, N. Marchal, A. Manzini, L. A. Hendricks, J. Mateos-Garcia, S. Bergman, J. Kay, C. Griffin, B. Bariach, I. Gabriel, V. Rieser and W. Isaac, Sociotechnical safety evaluation of generative AI systems, arXiv:2310.11986 (2023)."},{"key":"S1793351X26410011BIB024","unstructured":"European Parliament and Council of the European Union. Regulation (EU) 2024\/1689 of 13 June 2024 laying down harmonised rules on artificial intelligence (Artificial Intelligence Act). Official Journal of the European Union, OJ L, 2024\/1689, 12.7.2024. CELEX: 32024R1689."},{"key":"S1793351X26410011BIB025","volume-title":"IEEE Standard Model Process for Addressing Ethical Concerns during System Design","author":"IEEE Standards Association","year":"2021"},{"key":"S1793351X26410011BIB026","volume-title":"Artificial Intelligence Risk Management Framework (AI RMF 1.0) (NIST AI 100-1)","author":"National Institute of Standards and Technology","year":"2023"},{"key":"S1793351X26410011BIB027","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2025.uncertainlp-main.19"},{"key":"S1793351X26410011BIB028","doi-asserted-by":"publisher","DOI":"10.1145\/3458723"},{"key":"S1793351X26410011BIB029","first-page":"2148","volume":"26","author":"Denil M.","year":"2013","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"S1793351X26410011BIB030","unstructured":"M. Marion, A. \u00dcst\u00fcn, L. Pozzobon, A. Wang, M. Fadaee and S. Hooker, When less is more: Investigating data pruning for pretraining LLMs at scale, arXiv:2309.04564 (2023)."},{"key":"S1793351X26410011BIB031","first-page":"5998","volume":"30","author":"Vaswani","year":"2017","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"S1793351X26410011BIB032","volume-title":"Int. Conf. on Learning Representations","author":"Frankle J.","year":"2019"}],"container-title":["International Journal of Semantic Computing"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.worldscientific.com\/doi\/pdf\/10.1142\/S1793351X26410011","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,21]],"date-time":"2026-04-21T04:27:28Z","timestamp":1776745648000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.worldscientific.com\/doi\/10.1142\/S1793351X26410011"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,3]]},"references-count":32,"journal-issue":{"issue":"01","published-print":{"date-parts":[[2026,3]]}},"alternative-id":["10.1142\/S1793351X26410011"],"URL":"https:\/\/doi.org\/10.1142\/s1793351x26410011","relation":{},"ISSN":["1793-351X","1793-7108"],"issn-type":[{"value":"1793-351X","type":"print"},{"value":"1793-7108","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,3]]}}}