{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,2,5]],"date-time":"2026-02-05T07:33:01Z","timestamp":1770276781206,"version":"3.49.0"},"reference-count":20,"publisher":"Association for Computing Machinery (ACM)","issue":"9","license":[{"start":{"date-parts":[[2025,8,28]],"date-time":"2025-08-28T00:00:00Z","timestamp":1756339200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":["Commun. ACM"],"published-print":{"date-parts":[[2025,9]]},"abstract":"<jats:p>Three risks inherent in LLMs.<\/jats:p>","DOI":"10.1145\/3749447","type":"journal-article","created":{"date-parts":[[2025,8,19]],"date-time":"2025-08-19T18:01:24Z","timestamp":1755626484000},"page":"46-53","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":2,"title":["The Price of Intelligence"],"prefix":"10.1145","volume":"68","author":[{"given":"Mark","family":"Russinovich","sequence":"first","affiliation":[{"name":"Microsoft Azure, Bellevue, WA, USA"}]},{"given":"Ahmed","family":"Salem","sequence":"additional","affiliation":[{"name":"Microsoft Security Response Center, Redmond, WA, USA"}]},{"given":"Santiago","family":"Zanella-B\u00e9guelin","sequence":"additional","affiliation":[{"name":"Microsoft Research, Cambridge, United Kingdom"}]},{"given":"Yonatan","family":"Zunger","sequence":"additional","affiliation":[{"name":"Microsoft, Mountain View, CA, USA"}]}],"member":"320","published-online":{"date-parts":[[2025,8,28]]},"reference":[{"key":"e_1_3_1_2_2","doi-asserted-by":"publisher","DOI":"10.2196\/53164"},{"key":"e_1_3_1_3_2","unstructured":"Chern I.-C. et al. FacTool: Factuality detection in generative AI \u2014A tool augmented framework for multi-task and multi-domain scenarios. arXiv (2023); https:\/\/arxiv.org\/abs\/2307.13528."},{"key":"e_1_3_1_4_2","unstructured":"Forbes G. Levin E. and Beltagy I. Metric ensembles for hallucination detection.\u00a0arXiv (2023); https:\/\/arxiv.org\/abs\/2310.10495."},{"key":"e_1_3_1_5_2","doi-asserted-by":"publisher","DOI":"10.1145\/3571730"},{"key":"e_1_3_1_6_2","doi-asserted-by":"publisher","DOI":"10.1093\/jcmc\/zmac029"},{"key":"e_1_3_1_7_2","unstructured":"Lightman H. et al. Let\u2019s verify step by step. arXiv (2023); https:\/\/arxiv.org\/abs\/2305.20050."},{"key":"e_1_3_1_8_2","doi-asserted-by":"publisher","DOI":"10.1177\/0018720810376055"},{"key":"e_1_3_1_9_2","doi-asserted-by":"crossref","unstructured":"Robinette P. et al. Overtrust of robots in emergency evacuation scenarios. In Proceedings of the 11th ACM\/IEEE Intern. Conf. on Human-Robot Interaction (2016) 101\u2013108; https:\/\/dl.acm.org\/doi\/10.5555\/2906831.2906851.","DOI":"10.1109\/HRI.2016.7451740"},{"key":"e_1_3_1_10_2","unstructured":"Weller O. et al. FollowIR: Evaluating and teaching information retrieval models to follow instructions. (2024); https:\/\/arxiv.org\/abs\/2403.15246."},{"key":"e_1_3_1_11_2","doi-asserted-by":"publisher","DOI":"10.1177\/0018720815581940"},{"key":"e_1_3_1_12_2","unstructured":"Zhang T. et al. RAFT: Adapting language model to domain specific RAG. arXiv (2024); https:\/\/arxiv.org\/abs\/2403.10131."},{"key":"e_1_3_1_13_2","unstructured":"Gu et al. Exploring the role of instruction tuning in mitigating prompt injection attacks in large language model. (Claude 3.5 Sonnet hallucinated this reference. The paper does not exist; the link points to a paper on Astrophysics) (2023); https:\/\/arxiv.org\/abs\/2306.10783"},{"key":"e_1_3_1_14_2","unstructured":"Hines K. et al. Defending against indirect prompt injection attacks with spotlighting. arXiv (2024); https:\/\/arxiv.org\/abs\/2403.14720."},{"key":"e_1_3_1_15_2","unstructured":"Liu Y. et al. Prompt injection attack against LLM-integrated applications. arXiv (2023); https:\/\/arxiv.org\/abs\/2306.05499."},{"key":"e_1_3_1_16_2","unstructured":"Wallace E. et al. The instruction hierarchy: Training LLMs to prioritize privileged instructions. arXiv (2024); https:\/\/arxiv.org\/abs\/2404.13208."},{"key":"e_1_3_1_17_2","unstructured":"Russinovich M. Mitigating Skeleton Key a new type of generative AI jailbreak technique. Microsoft Security Blog (2024); https:\/\/www.microsoft.com\/en-us\/security\/blog\/2024\/06\/26\/mitigating-skeleton-key-a-new-type-of-generative-ai-jailbreak-technique\/."},{"key":"e_1_3_1_18_2","unstructured":"Russinovich M. Salem A. and Eldan R. Great now write an article about that: The Crescendo multi-turn LLM jailbreak attack. arXiv (2024); https:\/\/arxiv.org\/abs\/2404.01833."},{"key":"e_1_3_1_19_2","doi-asserted-by":"crossref","unstructured":"Shen X. et al. Do anything now: Characterizing and evaluating in-the-wild jailbreak prompts on large language models. In Proceedings of the 31st ACM SIGSAC Conf. on Computer and Communications Security (2024); https:\/\/arxiv.org\/abs\/2308.03825.","DOI":"10.1145\/3658644.3670388"},{"key":"e_1_3_1_20_2","unstructured":"Weatherbed J. Trolls have flooded X with graphic Taylor Swift AI fakes. The Verge (Jan. 25 2024); https:\/\/www.theverge.com\/2024\/1\/25\/24050334\/x-twitter-taylor-swift-ai-fake-images-trending."},{"key":"e_1_3_1_21_2","unstructured":"Zou A. et al. Universal and transferable adversarial attacks on aligned language models. arXiv (2023); https:\/\/arxiv.org\/abs\/2307.15043."}],"container-title":["Communications of the ACM"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3749447","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,8,28]],"date-time":"2025-08-28T13:55:55Z","timestamp":1756389355000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3749447"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,8,28]]},"references-count":20,"journal-issue":{"issue":"9","published-print":{"date-parts":[[2025,9]]}},"alternative-id":["10.1145\/3749447"],"URL":"https:\/\/doi.org\/10.1145\/3749447","relation":{},"ISSN":["0001-0782","1557-7317"],"issn-type":[{"value":"0001-0782","type":"print"},{"value":"1557-7317","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,8,28]]},"assertion":[{"value":"2025-08-28","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}