{"version":"1.0","type":"rich","provider_name":"techandbusiness.org","provider_url":"https://techandbusiness.org","title":"Study tests training control for role drift in multi-model AI systems","author_name":"techandbusiness.org · AI, Tech & Business","thumbnail_url":"https://images.ctfassets.net/jdtwqhzvc2n1/3JU47HcWZul3CRNEQLwMus/780888ab62ab547d48a4d3bbfdeeff72/Faulty_AI_pipeline.jpg?w=800&q=75","width":600,"height":400,"html":"<blockquote class=\"tb-newswire-embed\" style=\"max-width:600px;border-left:3px solid #22d3ee;padding:12px 16px;margin:0;font-family:-apple-system,system-ui,sans-serif;background:#09090b;border-radius:0 8px 8px 0;\">\n      <p style=\"margin:0 0 8px;font-size:10px;font-weight:600;letter-spacing:0.1em;color:#71717a;\">techandbusiness.org · AI, Tech & Business</p>\n      <p style=\"margin:0 0 8px;font-size:18px;font-weight:700;line-height:1.3;color:#fff;\"><a href=\"https://techandbusiness.org/newswire/GHLiZykD4nMQticyc98xp7\" style=\"color:#fff;text-decoration:none;\">Study tests training control for role drift in multi-model AI systems</a></p>\n      <p style=\"margin:0;font-size:14px;color:#a1a1aa;line-height:1.5;\">MIT and Harvard researchers report Role Anchor, a training method intended to keep modules in compound LLM pipelines aligned with assigned roles. In tests, an unanchored decomposer began inserting ans...</p>\n    </blockquote>"}