{"slug": "reading-is-not-using-retrieval-judgment-and-ai-financial-research", "title": "Reading Is Not Using: Retrieval, Judgment, and AI Financial Research", "summary": "A new arXiv preprint (2608.24842) finds that large language models (LLMs) used as AI analysts can retrieve risk disclosures from financial filings yet fail to incorporate them into investment judgments, with a risk disclosure's influence falling to the experimental noise floor even as direct retrieval remains accurate when unrelated context is varied from 2,000 to 128,000 tokens. The study, which replicates across model families and judgment tasks and in experiments with real 10-K filings, shows that workflow architecture—such as chunk-and-summarize pipelines versus targeted structured restatement—determines whether retrieved information affects judgments, implying that retrieval-based evaluations can certify systems whose investment decisions ignore information they demonstrably retrieved.", "body_md": "# Computer Science > Computation and Language\n\n[Submitted on 25 Aug 2026]\n\n# Title:Reading Is Not Using: Retrieval, Judgment, and the Design of AI Financial Research Workflows\n\n[View PDF](/pdf/2608.24842)\n\n[HTML (experimental)](https://arxiv.org/html/2608.24842v1)\n\nAbstract:Large language models (LLMs) are increasingly deployed as AI analysts to process financial disclosures and support AI-assisted investment decisions. Yet such systems are usually evaluated by what they can retrieve, not whether retrieved information affects their judgments. We identify a retrieval-integration gap in long-context financial analysis. Holding focal-firm information fixed and varying only unrelated context from 2,000 to 128,000 tokens, we find that a risk disclosure's influence on investment judgments falls to the experimental noise floor even as direct retrieval remains accurate. The pattern replicates across model families and judgment tasks and in experiments removing real disclosures from actual 10-K filings. More capable models postpone but do not eliminate the gap. Causal memory interventions show that compressed summaries and source-text lookup jointly transmit disclosures into judgments. Workflow architecture determines whether this transmission succeeds: chunk-and-summarize pipelines evict relevant information, whereas a targeted, structured restatement adjacent to the decision restores its influence. AI analyst performance is therefore jointly determined by model capability and workflow architecture. Retrieval-based evaluations can certify systems whose investment judgments ignore information they demonstrably retrieved.\n\n### References & Citations\n\nLoading...\n\n# Bibliographic and Citation Tools\n\nBibliographic Explorer\n\n*(*[What is the Explorer?](https://info.arxiv.org/labs/showcase.html#arxiv-bibliographic-explorer))\nConnected Papers\n\n*(*[What is Connected Papers?](https://www.connectedpapers.com/about))\nLitmaps\n\n*(*[What is Litmaps?](https://www.litmaps.co/))\nscite Smart Citations\n\n*(*[What are Smart Citations?](https://www.scite.ai/))# Code, Data and Media Associated with this Article\n\nalphaXiv\n\n*(*[What is alphaXiv?](https://alphaxiv.org/))\nCatalyzeX Code Finder for Papers\n\n*(*[What is CatalyzeX?](https://www.catalyzex.com))\nDagsHub\n\n*(*[What is DagsHub?](https://dagshub.com/))\nGotit.pub\n\n*(*[What is GotitPub?](http://gotit.pub/faq))\nHugging Face\n\n*(*[What is Huggingface?](https://huggingface.co/huggingface))\nScienceCast\n\n*(*[What is ScienceCast?](https://sciencecast.org/welcome))# Demos\n\n# Recommenders and Search Tools\n\nInfluence Flower\n\n*(*[What are Influence Flowers?](https://influencemap.cmlab.dev/))\nCORE Recommender\n\n*(*[What is CORE?](https://core.ac.uk/services/recommender))# arXivLabs: experimental projects with community collaborators\n\narXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.\n\nBoth individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.\n\nHave an idea for a project that will add value for arXiv's community? [ Learn more about arXivLabs](https://info.arxiv.org/labs/index.html).", "url": "https://wpnews.pro/news/reading-is-not-using-retrieval-judgment-and-ai-financial-research", "canonical_source": "https://arxiv.org/abs/2608.24842", "published_at": "2026-08-27 02:07:07+00:00", "updated_at": "2026-08-27 02:18:13.415086+00:00", "lang": "en", "topics": ["artificial-intelligence", "large-language-models", "ai-research"], "entities": ["arXiv"], "alternates": {"html": "https://wpnews.pro/news/reading-is-not-using-retrieval-judgment-and-ai-financial-research", "markdown": "https://wpnews.pro/news/reading-is-not-using-retrieval-judgment-and-ai-financial-research.md", "text": "https://wpnews.pro/news/reading-is-not-using-retrieval-judgment-and-ai-financial-research.txt", "jsonld": "https://wpnews.pro/news/reading-is-not-using-retrieval-judgment-and-ai-financial-research.jsonld"}}