{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,28]],"date-time":"2026-09-28T04:56:13Z","timestamp":1790571373053,"version":"4.1.0"},"reference-count":47,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,9,28]],"date-time":"2026-09-28T00:00:00Z","timestamp":1790553600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Big Data"],"abstract":"<jats:p>Large language model agents that persist across sessions, tools, users, and changing environments do more than answer isolated prompts; they accumulate state. When new evidence arrives, the central question is where that change should live: transient context, external memory, tool or workflow definitions, activation states, or model parameters. Recent benchmarks compare update mechanisms and assess memory over time, while emerging architectures coordinate multiple memory types. In this Perspective, we examine how a decision layer should select among update substrates and evaluate whether its choices remain appropriate. A controlled update is not merely a successful edit or a recalled fact; it is a decision to alter the least invasive substrate sufficient for the claim's scope, expected persistence, and evidentiary strength, including when the correct action is not to write persistent state. We propose a substrate-aware view of controlled knowledge updating organized by three principles: substrate proportionality, temporal defeasibility, and auditable continuity. This framing treats knowledge updating as a longitudinal control problem and motivates evaluation criteria that include update selection, temporal consistency, interference, reversibility, efficiency, and robustness. The resulting agenda connects memory, knowledge editing, agent evolution, and benchmarking under a single practical question: how should an agent decide what to change so that future behavior improves without accumulating uncontrolled state?<\/jats:p>","DOI":"10.3389\/fdata.2026.1938137","type":"journal-article","created":{"date-parts":[[2026,9,28]],"date-time":"2026-09-28T04:28:54Z","timestamp":1790569734000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Controlled knowledge updating in memory-enabled AI agents: beyond model editing and recall"],"prefix":"10.3389","volume":"9","author":[{"given":"Gabriel","family":"Chavira Ju\u00e1rez","sequence":"first","affiliation":[{"name":"Facultad de Ingenier\u00eda Tampico, Universidad Aut\u00f3noma de Tamaulipas","place":["Tampico, Tamaulipas, Mexico"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Eder Jahir","family":"Gonzalez Bravo","sequence":"additional","affiliation":[{"name":"Facultad de Ingenier\u00eda Tampico, Universidad Aut\u00f3noma de Tamaulipas","place":["Tampico, Tamaulipas, Mexico"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Guadalupe Esmeralda","family":"Rivera Garc\u00eda","sequence":"additional","affiliation":[{"name":"Tecnol\u00f3gico Nacional de M\u00e9xico\/Instituto Tecnol\u00f3gico Superior de P\u00e1nuco","place":["P\u00e1nuco, Veracruz, Mexico"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Javier A.","family":"Arcos Espinosa","sequence":"additional","affiliation":[{"name":"Facultad de Ingenier\u00eda Tampico, Universidad Aut\u00f3noma de Tamaulipas","place":["Tampico, Tamaulipas, Mexico"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Salvador W.","family":"Nava-D\u00edaz","sequence":"additional","affiliation":[{"name":"Facultad de Ingenier\u00eda Tampico, Universidad Aut\u00f3noma de Tamaulipas","place":["Tampico, Tamaulipas, Mexico"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,9,28]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"1749517","DOI":"10.3389\/frai.2026.1749517","article-title":"From simulated empathy to structural attunement: realtime editable memory topology and the evolution of emotionally grounded AI","volume":"9","author":"Albanese","year":"2026","journal-title":"Front. Artificial Intelligence"},{"key":"B2","article-title":"\u201cSelf-RAG: learning to retrieve, generate, and critique through self-reflection,\u201d","author":"Asai","year":"2024","journal-title":"International conference on learning representations"},{"key":"B3","first-page":"3119","article-title":"\u201cLongBench: a bilingual, multitask benchmark for long context understanding,\u201d","volume-title":"Proceedings of the 62nd annual meeting of the association for computational linguistics","author":"Bai","year":"2024"},{"key":"B4","first-page":"2206","article-title":"\u201cImproving language models by retrieving from trillions of tokens,\u201d","volume-title":"Proceedings of the 39th international conference on machine learning","author":"Borgeaud","year":"2022"},{"key":"B5","doi-asserted-by":"crossref","first-page":"141","DOI":"10.1109\/SP40001.2021.00019","article-title":"\u201cMachine unlearning,\u201d","volume-title":"2021 IEEE symposium on security and privacy (SP)","author":"Bourtoule","year":"2021"},{"key":"B6","doi-asserted-by":"crossref","first-page":"463","DOI":"10.1109\/SP.2015.35","article-title":"\u201cTowards making systems forget with machine unlearning,\u201d","volume-title":"2015 IEEE Symposium on Security and Privacy","author":"Cao","year":"2015"},{"key":"B7","doi-asserted-by":"crossref","first-page":"2978","DOI":"10.18653\/v1\/P19-1285","article-title":"\u201cTransformer-XL: attentive language models beyond a fixed-length context,\u201d","volume-title":"Proceedings of the 57th annual meeting of the association for computational linguistics","author":"Dai","year":"2019"},{"key":"B8","doi-asserted-by":"crossref","first-page":"6491","DOI":"10.18653\/v1\/2021.emnlp-main.522","article-title":"\u201cEditing factual knowledge in language models,\u201d","volume-title":"Proceedings of the 2021 conference on empirical methods in natural language processing","author":"De Cao","year":"2021"},{"key":"B9","doi-asserted-by":"publisher","first-page":"28091","DOI":"10.52202\/075280-1220","article-title":"\u201cMind2Web: towards a generalist agent for the web,\u201d","author":"Deng","year":"2023","journal-title":"Advances in neural information processing systems"},{"key":"B10","first-page":"150","article-title":"\u201cRAGAs: automated evaluation of retrieval augmented generation,\u201d","volume-title":"Proceedings of the 18th conference of the european chapter of the association for computational linguistics: system demonstrations","author":"Es","year":"2024"},{"key":"B11","doi-asserted-by":"crossref","first-page":"79","DOI":"10.1145\/3605764.3623985","article-title":"\u201cNot what you've signed up for: compromising real-world LLM-integrated applications with indirect prompt injection,\u201d","volume-title":"Proceedings of the 16th ACM workshop on artificial intelligence and security","author":"Greshake","year":"2023"},{"key":"B12","first-page":"3929","article-title":"\u201cRetrieval augmented language model pre-training,\u201d","volume-title":"Proceedings of the 37th international conference on machine learning","author":"Guu","year":"2020"},{"key":"B13","article-title":"\u201cEvaluating memory in LLM agents via incremental multi-turn interactions,\u201d","author":"Hu","year":"2026","journal-title":"The fourteenth international conference on learning representations"},{"key":"B14","article-title":"\u201cSWE-bench: can language models resolve real-world GitHub issues?,\u201d","author":"Jimenez","year":"2024","journal-title":"International conference on learning representations"},{"key":"B15","doi-asserted-by":"crossref","first-page":"25961","DOI":"10.18653\/v1\/2025.emnlp-main.1318","article-title":"\u201cMemory OS of AI agent,\u201d","volume-title":"Proceedings of the 2025 conference on empirical methods in natural language processing","author":"Kang","year":"2025"},{"key":"B16","first-page":"6769","article-title":"\u201cDense passage retrieval for open-domain question answering,\u201d","volume-title":"Proceedings of the 2020 conference on empirical methods in natural language processing","author":"Karpukhin","year":"2020"},{"key":"B17","doi-asserted-by":"publisher","first-page":"3521","DOI":"10.1073\/pnas.1611835114","article-title":"Overcoming catastrophic forgetting in neural networks","volume":"114","author":"Kirkpatrick","year":"2017","journal-title":"Proc. Nat. Acad. Sci"},{"key":"B18","first-page":"9459","article-title":"Retrieval-augmented generation for knowledge-intensive NLP tasks","volume":"33","author":"Lewis","year":"2020","journal-title":"Adv. Neural Information Proc. Syst"},{"key":"B19","doi-asserted-by":"publisher","first-page":"41451","DOI":"10.52202\/075280-1797","article-title":"Inference-time intervention: eliciting truthful answers from a language model","volume":"36","author":"Li","year":"2023","journal-title":"Adv. Neural Information Proc. Syst"},{"key":"B20","doi-asserted-by":"publisher","first-page":"3102","DOI":"10.18653\/v1\/2023.emnlp-main.187","article-title":"\u201cAPI-bank: a comprehensive benchmark for tool-augmented LLMs,\u201d","author":"Li","year":"2023","journal-title":"Proceedings of the 2023 conference on empirical methods in natural language processing"},{"key":"B21","doi-asserted-by":"publisher","first-page":"30850","DOI":"10.52202\/079017-0971","article-title":"Should we really edit language models? on the evaluation of edited language models","volume":"37","author":"Li","year":"2024","journal-title":"Adv. Neural Info. Proc. Syst"},{"key":"B22","article-title":"MemOS: a memory OS for AI system","author":"Li","year":"2025","journal-title":"arXiv"},{"key":"B23","doi-asserted-by":"publisher","first-page":"157","DOI":"10.1162\/tacl_a_00638","article-title":"Lost in the middle: how language models use long contexts","volume":"12","author":"Liu","year":"2024","journal-title":"Trans. Association Comput. Linguistics"},{"key":"B24","article-title":"\u201cAgentBench: evaluating LLMs as agents,\u201d","author":"Liu","year":"2024","journal-title":"International conference on learning representations"},{"key":"B25","first-page":"13851","article-title":"\u201cEvaluating very long-term conversational memory of LLM agents,\u201d","volume-title":"Proceedings of the 62nd annual meeting of the association for computational linguistics","author":"Maharana","year":"2024"},{"key":"B26","doi-asserted-by":"publisher","first-page":"17359","DOI":"10.52202\/068431-1262","article-title":"Locating and editing factual associations in GPT","volume":"35","author":"Meng","year":"2022","journal-title":"Adv. Neural Informat. Proc. Syst"},{"key":"B27","article-title":"\u201cMass-editing memory in a transformer,\u201d","author":"Meng","year":"2023","journal-title":"International conference on learning representations"},{"key":"B28","article-title":"\u201cFast model editing at scale,\u201d","author":"Mitchell","year":"","journal-title":"International conference on learning representations"},{"key":"B29","first-page":"15817","article-title":"\u201cMemory-based model editing at scale,\u201d","volume-title":"Proceedings of the 39th international conference on machine learning","author":"Mitchell","year":""},{"key":"B30","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1145\/3586183.3606763","article-title":"\u201cGenerative agents: interactive simulacra of human behavior,\u201d","volume-title":"Proceedings of the 36th annual ACM symposium on user interface software and technology","author":"Park","year":"2023"},{"key":"B31","doi-asserted-by":"publisher","first-page":"126544","DOI":"10.52202\/079017-4020","article-title":"Gorilla: large language model connected with massive APIs","volume":"37","author":"Patil","year":"2024","journal-title":"Adv. Neural Informat. Proc. Syst"},{"key":"B32","article-title":"ToolLLM: facilitating large language models to master 16000+ real-world APIs","author":"Qin","year":"2024","journal-title":"International conference on learning representations"},{"key":"B33","doi-asserted-by":"publisher","first-page":"68539","DOI":"10.52202\/075280-2997","article-title":"Toolformer: language models can teach themselves to use tools","volume":"36","author":"Schick","year":"2023","journal-title":"Adv. Neural Info. Proc. Syst"},{"key":"B34","doi-asserted-by":"publisher","first-page":"8634","DOI":"10.52202\/075280-0377","article-title":"Reflexion: language agents with verbal reinforcement learning","volume":"36","author":"Shinn","year":"2023","journal-title":"Adv. Neural Info. Proc. Syst"},{"key":"B35","first-page":"59326","article-title":"\u201cWikiBigEdit: understanding the limits of lifelong knowledge editing in LLMs,\u201d","volume-title":"Proceedings of the 42nd international conference on machine learning","author":"Thede","year":"2025"},{"key":"B36","doi-asserted-by":"publisher","first-page":"186345","DOI":"10.1007\/s11704-024-40231-1","article-title":"A survey on large language model based autonomous agents","volume":"18","author":"Wang","year":"2024","journal-title":"Front. Comput. Sci"},{"key":"B37","doi-asserted-by":"crossref","first-page":"7097","DOI":"10.18653\/v1\/2024.findings-emnlp.416","article-title":"\u201cKnowledge mechanisms in large language models: a survey and perspective,\u201d","volume-title":"Findings of the association for computational linguistics: EMNLP 2024","author":"Wang","year":"2024"},{"key":"B38","first-page":"82","article-title":"\u201cEasyEdit: an easy-to-use knowledge editing framework for large language models,\u201d","volume-title":"Proceedings of the 62nd annual meeting of the association for computational linguistics","author":"Wang","year":"2024"},{"key":"B39","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1145\/3698590","article-title":"Knowledge editing for large language models: a survey","volume":"57","author":"Wang","year":"2025","journal-title":"ACM Computing Surveys"},{"key":"B40","article-title":"\u201cMemorizing transformers,\u201d","author":"Wu","year":"2022","journal-title":"International conference on learning representations"},{"key":"B41","first-page":"17577","article-title":"\u201cA-Mem: agentic memory for LLM agents,\u201d","volume-title":"Advances in Neural Information Processing Systems","author":"Xu","year":"2025"},{"key":"B42","doi-asserted-by":"publisher","first-page":"20744","DOI":"10.52202\/068431-1508","article-title":"\u201cWebShop: towards scalable real-world web interaction with grounded language agents,\u201d","volume":"35","author":"Yao","year":"2022","journal-title":"Adv. Neural Info. Proc. Syst"},{"key":"B43","article-title":"\u201cReAct: synergizing reasoning and acting in language models,\u201d","author":"Yao","year":"2023","journal-title":"Int. Conference Learning Representations"},{"key":"B44","first-page":"21457","article-title":"\u201cAgentic memory: learning unified long-term and short-term memory management for large language model agents,\u201d","volume-title":"Proceedings of the 64th annual meeting of the association for computational linguistics","author":"Yu","year":"2026"},{"key":"B45","doi-asserted-by":"publisher","first-page":"19724","DOI":"10.1609\/aaai.v38i17.29946","article-title":"\u201cMemorybank: enhancing large language models with long-term memory,\u201d","volume":"38","author":"Zhong","year":"2024","journal-title":"Proceedings of the AAAI conference on artificial intelligence"},{"key":"B46","doi-asserted-by":"crossref","first-page":"13764","DOI":"10.18653\/v1\/2023.emnlp-main.849","article-title":"\u201cPoisoning retrieval corpora by injecting adversarial passages,\u201d","volume-title":"Proceedings of the 2023 conference on empirical methods in natural language processing","author":"Zhong","year":"2023"},{"key":"B47","article-title":"WebArena: a realistic web environment for building autonomous agents","author":"Zhou","year":"2024","journal-title":"International conference on learning representations"}],"container-title":["Frontiers in Big Data"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdata.2026.1938137\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,28]],"date-time":"2026-09-28T04:28:59Z","timestamp":1790569739000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdata.2026.1938137\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,28]]},"references-count":47,"alternative-id":["10.3389\/fdata.2026.1938137"],"URL":"https:\/\/doi.org\/10.3389\/fdata.2026.1938137","relation":{},"ISSN":["2624-909X"],"issn-type":[{"value":"2624-909X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,28]]},"article-number":"1938137"}}