{"version":"https://jsonfeed.org/version/1.1","title":"Science briefings","description":"Research and science stories that affect builders, labs, product teams, and evidence-led technology decisions.","home_page_url":"https://guidances.org/topic/science","feed_url":"https://guidances.org/topic/science/feed.json","language":"en","mock_mode":false,"items":[{"id":"75cef9ef-1f35-4144-9e86-90b383fb4065","url":"https://guidances.org/en/article/75cef9ef-1f35-4144-9e86-90b383fb4065/fine-tuning-llms-for-single-atom-catalyst-design-where-ai-chemistry-meet-36eb497a","title":"Fine-Tuning LLMs for Single-Atom Catalyst Design: What the Research Establishes and What Remains to Be Verified","summary":"Researchers from IBM Research, EPFL, and ETH Zurich have fine-tuned a Granite-based large language model on nearly 3,000 single-atom catalyst publications to generate synthesis protocols from user-defined prompts. The study illustrates a potential role for LLMs in materials science R&D pipelines and raises questions about scientific AI platforms, laboratory automation, and regulatory context. Commercial deployment pathways and quantitative performance metrics are not verifiable from the available source alone.","date_published":"2026-06-21T00:09:04.46+00:00","date_modified":"2026-07-02T18:03:28.129383+00:00","tags":["science","single-atom catalyst AI","LLM fine-tuning materials science","Granite model chemistry","IBM AI research","heterogeneous catalyst design","scientific AI infrastructure","AI-assisted synthesis","materials discovery AI","domain-specific LLM","clean energy AI","단일 원자 촉매 AI","LLM 미세조정 소재 과학","과학 AI 인프라","IBM Granite 모델","AI 합성 프로토콜"],"external_url":"https://www.nature.com/articles/s42004-026-02046-y"},{"id":"c92ba364-6c0a-47ae-b958-809ac37bf529","url":"https://guidances.org/en/article/c92ba364-6c0a-47ae-b958-809ac37bf529/aws-summit-new-york-2026-how-amazon-is-rebuilding-its-enterprise-cloud-a-6ed78cc9","title":"AWS Summit New York 2026: How Amazon Is Rebuilding Its Enterprise Cloud Around AI Agents","summary":"At AWS Summit New York 2026, Amazon Web Services announced a coordinated set of agent-focused products—including AWS Continuum, AWS Context, and expanded Bedrock AgentCore capabilities—highlighting a platform direction that places AI agents at the center of enterprise software delivery.","date_published":"2026-06-20T10:54:28.568+00:00","date_modified":"2026-07-18T20:11:51.542364+00:00","tags":["science","AWS Summit New York 2026","AWS Continuum AI security","AWS Context knowledge graph","Amazon Bedrock AgentCore","enterprise AI agents","AWS DevOps Agent release management","AI agent infrastructure cloud","Amazon AWS enterprise strategy","agentic AI platform","cloud AI governance compliance","AWS enterprise software 2026","AI native security service","AMZN cloud revenue growth","enterprise AI deployment architecture"],"external_url":"https://www.aboutamazon.com/news/aws/aws-summit-nyc-2026-ai-agents"},{"id":"cbdee9b4-5e9d-4592-8b3a-c134f1635358","url":"https://guidances.org/en/article/cbdee9b4-5e9d-4592-8b3a-c134f1635358/microsoft-and-publicis-research-points-to-conversational-ai-search-as-an-7c8b3d37","title":"Microsoft and Publicis Research Points to Conversational AI Search as an Emerging Advertising Channel","summary":"A joint whitepaper from Microsoft Advertising and Publicis Groupe, published June 18, 2025, finds that 75% of users report equivalent or better satisfaction with conversational AI search versus traditional search, and that 46% of consumers who notice ads in that environment report an improved experience. The findings are a reference point for discussions about search advertising and Microsoft's ad product strategy.","date_published":"2026-06-18T05:43:32.389+00:00","date_modified":"2026-07-18T21:57:38.176791+00:00","tags":["science","conversational AI search advertising","Microsoft Advertising Publicis whitepaper","AI search consumer satisfaction","Copilot advertising revenue","AI native search monetization","digital advertising market 2025","Bing AI advertising","search advertising disruption","generative AI ad formats","Microsoft MSFT advertising strategy"],"external_url":"https://about.ads.microsoft.com/en/resources/discover/insights/publicis-research"},{"id":"75daa3cd-2860-4bbe-8355-d71b4983ab49","url":"https://guidances.org/en/article/75daa3cd-2860-4bbe-8355-d71b4983ab49/global-ai-leadership-perceptions-tilt-toward-china-raising-policy-and-ma-f80087e6","title":"Global AI Leadership Perceptions Shift Toward China, Raising Policy and Market Questions","summary":"A Public First survey of more than 18,000 respondents across 15 countries suggests that people in key U.S.-allied markets increasingly view China as the world’s AI leader, while American confidence in AI is weakening over resource use, labor displacement, and information reliability. The result matters as a signal that perception can influence procurement, regulation, and go-to-market strategy.","date_published":"2026-06-17T16:50:21.523+00:00","date_modified":"2026-07-18T19:18:46.854227+00:00","tags":["science","AI leadership perception","China AI leader","US AI policy","AI procurement South Korea Japan","AI regulation transparency","AI infrastructure capex","semiconductor export controls","enterprise AI trust","AI market sentiment","builder implications AI"],"external_url":"https://www.politico.com/news/2026/06/15/people-around-the-world-see-a-winner-on-ai-and-its-not-the-us-00960930"},{"id":"806e00b2-0ec0-404e-ac4e-e95bf821a477","url":"https://guidances.org/en/article/806e00b2-0ec0-404e-ac4e-e95bf821a477/deepmind-measures-ai-learning-effects-in-sierra-leone-school-trial-5892df4d","title":"DeepMind measures AI learning effects in Sierra Leone school trial","summary":"Google DeepMind says a randomized controlled trial across 12 schools in Sierra Leone and 1,763 junior secondary students found that guided AI learning lifted mathematics scores by 0.258 standard deviations. The result reinforces a broader shift in edtech: AI tools will increasingly be judged by learning outcomes, not by novelty or usage alone.","date_published":"2026-06-14T09:26:40.467+00:00","date_modified":"2026-08-13T02:38:39.785977+00:00","tags":["science","DeepMind","AI education","randomized controlled trial","Sierra Leone","edtech","guided learning","math scores","learning outcomes"],"external_url":"https://deepmind.google/blog/measuring-the-impact-of-learning-with-ai-in-sierra-leone-and-beyond"},{"id":"be13ee8a-020c-4f26-aa01-4143a59bc0d5","url":"https://guidances.org/en/article/be13ee8a-020c-4f26-aa01-4143a59bc0d5/stanford-advances-real-time-clinical-validation-studies-for-medical-imag-487ddbde","title":"Stanford Advances Real-Time Clinical Validation Studies for Medical Imaging AI Models","summary":"Stanford University's Center for Artificial Intelligence in Medicine & Imaging is conducting prospective real-time clinical validation studies of AI models for medical imaging. This is a systematic approach to evaluating the safety and effectiveness of AI tools in actual clinical settings, helping build evidence that can inform regulatory review and healthcare deployment.","date_published":"2026-06-14T03:44:04.038+00:00","date_modified":"2026-07-08T18:31:27.92668+00:00","tags":["science","Stanford medical AI validation","clinical validation AI imaging","prospective AI studies healthcare","medical imaging AI deployment","FDA AI medical device approval","real-time clinical AI testing","healthcare AI evidence","radiology AI validation","AI clinical trials","medical AI regulatory approval","AI in medicine","healthcare technology validation","AI model robustness","clinical workflow integration AI"],"external_url":"https://aimi.stanford.edu/research/focal-areas/clinical-validation"},{"id":"d203e796-cc0a-41ec-aece-394b76c3a503","url":"https://guidances.org/en/article/d203e796-cc0a-41ec-aece-394b76c3a503/expert-level-academic-question-benchmark-offers-new-standard-for-ai-asse-5682a0ad","title":"Expert-Level Academic Question Benchmark Offers New Standard for AI Assessment","summary":"Nature has introduced a benchmark of expert-level academic questions designed to assess the scholarly capabilities of AI systems. The benchmark aims to move beyond existing evaluation tools by testing advanced reasoning abilities required in real research environments. The research community anticipates this will enable more accurate measurement of AI models' scientific problem-solving capacity.","date_published":"2026-06-14T02:16:21.47+00:00","date_modified":"2026-06-28T10:57:29.613113+00:00","tags":["science","AI benchmark","expert-level academic questions","AI assessment","Nature journal","Lab Bench","scientific reasoning AI","AI evaluation methodology","research AI tools","domain-specific AI","academic AI capabilities"],"external_url":"https://www.nature.com/articles/s41586-025-09962-4"},{"id":"662ed8c5-56fb-4187-8c21-e1c082f34941","url":"https://guidances.org/en/article/662ed8c5-56fb-4187-8c21-e1c082f34941/anthropic-calls-for-agent-friendly-infrastructure-in-biological-research-8380cbb9","title":"Anthropic Proposes Agent-Friendly Infrastructure for Biological Research","summary":"Anthropic has published a research blog post proposing that biological data infrastructure become more agent-friendly. The company outlines deterministic execution layers, reliable access to biological databases, and agent-accessible context engines to support scientific discovery.","date_published":"2026-06-12T07:25:39.464+00:00","date_modified":"2026-07-03T02:36:02.740202+00:00","tags":["science","Anthropic","AI agents","biological research","scientific discovery","data infrastructure","deterministic execution","biological databases","context engines","life sciences AI","research automation"],"external_url":"https://www.anthropic.com/research/agents-in-biology"},{"id":"231734a3-cd28-400f-883a-0243c6f4ab74","url":"https://guidances.org/en/article/231734a3-cd28-400f-883a-0243c6f4ab74/openai-introduces-paperbench-benchmark-to-evaluate-ai-research-replicati-9bd42a8d","title":"OpenAI Introduces PaperBench Benchmark to Evaluate AI Research Replication Capability","summary":"OpenAI has released PaperBench, a new benchmark designed to measure AI agents' ability to replicate state-of-the-art research. The benchmark evaluates how accurately AI systems can reproduce empirical contributions from published papers, establishing a new standard for automated scientific research capabilities.","date_published":"2026-06-12T04:44:08.433+00:00","date_modified":"2026-08-14T05:11:08.82481+00:00","tags":["science","PaperBench","OpenAI","AI research replication","research reproducibility","AI benchmark","scientific automation","research automation","AI agents","machine learning reproducibility","automated research verification"],"external_url":"https://cdn.openai.com/papers/22265bac-3191-44e5-b057-7aaacd8e90cd/paperbench.pdf"}]}