[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-ai-security-research-keeps-breaking-things-not-fixing-them":10,"sections":35},{"siteName":4,"siteTagline":5,"publisherName":4,"contactEmail":6},"The Revision","Tech news, decoded.","editor@therevision.news",{"gaMeasurementId":8,"adsenseClientId":9},"G-ZW2MV82GYR","ca-pub-8533917693782264",{"article":11},{"id":12,"slug":13,"title":14,"dek":15,"body_md":16,"tags_json":17,"published_at":18,"created_at":19,"updated_at":20,"status":21,"review_note":22,"review_notes":23,"image_url":22,"persona_id":22,"persona_name":22,"section":24,"tags":25,"sources":30,"feedback":34,"feedback_at":22,"cost_usd":34,"total_tokens":34},9305,"ai-security-research-keeps-breaking-things-not-fixing-them","AI Security Research Keeps Breaking Things, Not Fixing Them","A new paper finds AI security research rewards flashy attacks over working defenses, skewing incentives away from real protection.","AI security research is great at breaking things and bad at fixing them, according to a new review of the field.\n\nThe analysis looked across subfields including federated learning, speech recognition, membership inference, and large language models, and found a consistent skew: far more papers propose new attacks than propose defenses. Worse, the imbalance isn't just about volume. Attack papers tend to get evaluated under favorable conditions that make the threats look scarier than they'd be in the real world, while defense papers are judged against a much stricter bar. Few defenses clear that bar, so the literature piles up with demonstrated vulnerabilities and comes up short on protections anyone actually deploys.\n\nThat's a problem for anyone building or securing AI systems, because the research base that's supposed to inform real-world hardening is lopsided toward hype about what can go wrong rather than guidance on how to stop it. The authors' prescription is straightforward: evaluate defenses on fairer terms and give researchers more reason to work on them in the first place.\n\nIt's the academic equivalent of a locksmith who only picks locks and never fixes one. Useful to know about. Less useful if you're trying to secure the door.","[\"ai-security\",\"cybersecurity\",\"research\",\"academia\"]","2026-10-01T04:00:00.000Z","2026-10-02T08:15:07.542Z","2026-10-02T08:15:09.430Z","published",null,[],"security",[26,27,28,29],"ai-security","cybersecurity","research","academia",[31],{"name":32,"url":33},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2605.23448",0,{"sections":36},[37,41,44,48,53,58,62,67,72,76,81,86,91,96],{"name":38,"slug":39,"count":40,"latest_published_at":18},"AI","ai",5688,{"name":42,"slug":24,"count":43,"latest_published_at":18},"Security",820,{"name":45,"slug":46,"count":47,"latest_published_at":18},"Policy","policy",430,{"name":49,"slug":50,"count":51,"latest_published_at":52},"Deals","deals",302,"2026-10-01T09:05:03.000Z",{"name":54,"slug":55,"count":56,"latest_published_at":57},"Hardware","hardware",196,"2026-09-30T13:00:00.000Z",{"name":59,"slug":60,"count":61,"latest_published_at":18},"Science","science",165,{"name":63,"slug":64,"count":65,"latest_published_at":66},"Consumer Tech","consumer-tech",149,"2026-09-30T22:57:11.000Z",{"name":68,"slug":69,"count":70,"latest_published_at":71},"Dev Tools","dev-tools",93,"2026-10-01T02:30:48.000Z",{"name":73,"slug":74,"count":70,"latest_published_at":75},"Software","software","2026-09-30T21:41:11.000Z",{"name":77,"slug":78,"count":79,"latest_published_at":80},"Startups","startups",84,"2026-09-30T20:39:09.000Z",{"name":82,"slug":83,"count":84,"latest_published_at":85},"Gaming","gaming",51,"2026-09-30T16:24:30.000Z",{"name":87,"slug":88,"count":89,"latest_published_at":90},"General","general",50,"2026-09-30T21:37:54.000Z",{"name":92,"slug":93,"count":94,"latest_published_at":95},"Reviews","reviews",31,"2026-09-28T14:31:34.000Z",{"name":97,"slug":98,"count":99,"latest_published_at":100},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]