{"paper":{"title":"Strategic commitments shape collective cybersecurity under AI inequality","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"Subsidies for a small group of committed defenders can spread strong protection and cut successful attacks even when AI tools are costly for most.","cross_cats":["physics.soc-ph"],"primary_cat":"cs.AI","authors_text":"Adeela Bashir, Matjaz Perc, The Anh Han, Zhao Song, Zia Ush Shamszaman","submitted_at":"2026-05-10T08:34:51Z","abstract_excerpt":"The growing integration of AI into cybersecurity is reshaping the balance between attackers and defenders. When access to advanced AI-enabled defence tools is uneven, resource-limited defenders may be unable to adopt effective protection, creating persistent system vulnerabilities. We study the impact of differential AI access using an evolutionary game-theoretic model in a finite population. We first show that when high-capability defence is costly, the population is driven toward low-cost, weak-defence behaviour, sustaining attacks and weakening long-run security. To address this problem, we"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Our analysis shows that subsidised commitment significantly increases strong defence adoption, suppresses successful attacks, and improves overall system resilience.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The model assumes that social learning through imitation allows committed defenders to influence the broader population and that a targeted subsidy can be applied without triggering new strategic responses or implementation costs.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Subsidized commitment by a small group of defenders in an evolutionary game model significantly increases strong defense adoption, suppresses attacks, and improves system resilience under AI access inequality.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Subsidies for a small group of committed defenders can spread strong protection and cut successful attacks even when AI tools are costly for most.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"5434c3c7296ac3942014ed78385e66325399bd0928db7fe61bb4b65b6537d01b"},"source":{"id":"2605.09415","kind":"arxiv","version":2},"verdict":{"id":"eb56f983-6ef7-4e75-acc6-9306e2d6e8a1","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T21:32:05.163212Z","strongest_claim":"Our analysis shows that subsidised commitment significantly increases strong defence adoption, suppresses successful attacks, and improves overall system resilience.","one_line_summary":"Subsidized commitment by a small group of defenders in an evolutionary game model significantly increases strong defense adoption, suppresses attacks, and improves system resilience under AI access inequality.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The model assumes that social learning through imitation allows committed defenders to influence the broader population and that a targeted subsidy can be applied without triggering new strategic responses or implementation costs.","pith_extraction_headline":"Subsidies for a small group of committed defenders can spread strong protection and cut successful attacks even when AI tools are costly for most."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.09415/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"claim_evidence","ran_at":"2026-05-20T07:42:01.466372Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T19:35:55.040268Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T13:01:18.257551Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T10:17:44.738011Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"dd7e271370d1228137e7487a7b7f3281253e0d5739f421afd1be5cc8a8df129d"},"references":{"count":63,"sample":[{"doi":"","year":2014,"title":"Artificial (intelligent) agents and active cyber defence: Policy implications","work_id":"bcaf49a2-f919-4d39-a301-f9c67241cb54","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2020,"title":"Artificial intelligence in the cyber domain: Offense and defense.Symmetry, 12(3):410","work_id":"d609f500-dc6e-4f5b-8827-200a53fbe4d6","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2021,"title":"A survey on adversarial attacks and defences.CAAI Transactions on In- telligence Technology, 6(1):25–45","work_id":"854ddff6-ae90-48d6-a8cb-e8ace3cf8253","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Leveraging ai/ml for anomaly detection, threat prediction, and au- tomated response.World Journal of Advanced Research and Reviews","work_id":"af735832-0629-477b-a04f-7e86f8d19af8","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Ai-powered cyberattacks: A comprehensive review and analysis of emerging threats.Advances in IT and Electrical Engineering, 31:55–70","work_id":"eddf8586-024c-4492-b24c-27fb5b673baa","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":63,"snapshot_sha256":"c5ccf3b2447c2de72064394d28f3140350d19b99a5f0c7828e8ba45aa1c4f6c3","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}