This pasta sauce wants to record your family Even dinner wants to listen to your conversations now.

Why.

A new TikTok AI toggle is causing panic among creators TikTok creators say a new AI remix setting is enabled on videos by default, forcing users to manually review and change permissions.

Larry Ellison’s TikTok forcing “AI Remixing” onto their entire user base without their knowledge or consent is a really great microcosm of what US tech overlordship looks like — and why the backlash against this technology is only growing.

The Guardian view on social science research: embracing uncertainty | Editorial Editorial: Science rarely produces identical outcomes. Mistaking this for failure turns caution into an excuse for inaction

🗞️ A recent editorial in The Guardian draws on SCORE findings to examine how replication results are interpreted: “Not every failed replication signals a crisis...Results that don’t consistently replicate should be weighed against a wider evidence base when guiding policy”

Q&A with Alisha Bruton: Open Science in Integrative Medicine Research Research scientist and biostatistician Alisha Bruton shares how she integrates open science practices into her workflow, navigates data sharing decisions, and applies FAIR principles to make research findable and reusable over time.

In our latest researcher Q&A, research scientist and biostatistician Alisha Bruton shares how she integrates open science into her workflow, navigates data sharing in collaborative research, and applies FAIR principles to help make research more findable and reusable.

💡

Virtual Event April 16 // 1 pm ET NEW EVIDENCE ON REPRODUCIBILITY ACROSS SOCIAL AND BEHAVIORAL RESEARCH Moderator: Tim Errington Speakers: Katrin Auspurg, Abel Brodeur, and Andrew Tyner

What can large-scale studies tell us about reproducibility? In our webinar on April 16, researchers from COS, I4R, and META-REP will discuss findings from three papers—one from the recently published SCORE effort—and insights on reproducibility, transparency, and credibility

cos-io.zoom.us/webin...

damn aaron swartz tried to warn everybody about sam altman

New on the Generalist Repository Ecosystem Initiative (GREI) blog: highlights from the Streamlining Data Sharing webinar series, featuring practical guidance and tools, real-world user stories, solutions to common data sharing challenges, and more!

📖 medium.com/@blog-gre...

The Center for Open Science Welcomes Chris Bourg and Marcus Munafò to its Board of Directors The Center for Open Science is pleased to welcome Chris Bourg (Massachusetts Institute of Technology) and Marcus Munafò (University of Bath) to the COS Board of Directors in 2026.

COS is pleased to welcome Chris Bourg (MIT) and Marcus Munafò (University of Bath) to our Board of Directors. Their leadership in equitable open scholarship, research culture reform, and metascience will help shape how the next phase of our work unfolds.

🎉

Welcome! You are invited to join a webinar: TOP 101: An Overview of the Transparency and Openness Promotion Guidelines. After registering, you will receive a confirmation email about joining the webinar. The Transparency and Openness Promotion Guidelines (TOP) is a policy framework that can provide specific recommendations for journals, research funders, universities, and researchers about practices that are designed to increase the verifiability of empirical research claims. TOP was updated in 2025 and contains three main parts: Research Practices, Verification Practices, and Verification Studies. This webinar will include an overview of these components and how TOP’s tiered and modular structure makes it applicable to a diverse community of researchers and policymakers. This webinar is part of COS’s ongoing campaign to demonstrate implementations of our Transparency and Openness Promotion (TOP) guidelines (https://bit.ly/top-guidelines), grounding our policy framework in real-world examples. If you have feedback, implementation stories, or lessons learned from working with TOP, get in touch with us at top@cos.io.

📋 TOP 101: An Overview of the Transparency & Openness Promotion Guidelines
April 1, 11 AM ET

Learn about the TOP framework, which provides recommendations for journals, funders, universities, & researchers about practices that can increase verifiability of research claims.

The 2025 COS Impact Report is now live! From open infrastructure to research initiatives, policy engagement, and global partnerships, see what COS accomplished over the past year to advance transparent, rigorous, and accessible research.

🔍 Check out the report: www.cos.io/impact

⏰ REMINDER: Applications for AREN's Local Network Leads program are open through March 16.

This program trains researchers & research professionals in #Africa to become #openscience leaders who can establish a community of practice at their institution.

Learn more & apply:

Building a Publishing Model for Replication: Q&A with the Senior Editors of Replication Research COS spoke with the senior editors of Replication Research—a community-led Diamond Open Access journal that supports reproduction and replication studies.

Replication Research (R2), a 🆕 community-led Diamond OA journal, makes replication studies more discoverable, publishable & rigorously evaluated—without subscription barriers or author fees. Ahead of #LoveReplicationsWeek, R2's senior editors shared their vision in our Q&A:

The updated Generalist Repository Ecosystem Initiative (GREI) flowchart offers clearer guidance and a more streamlined decision pathway to help researchers, librarians, and support teams choose the most appropriate generalist repository for their data.

📖 medium.com/@blog-gre...

🚀 NASA Summer Internship: Open Science Impact

This summer internship offers an opportunity to work with a NASA mentor on efforts to assess how open science practices accelerate discovery & broaden participation in science.

Deadline: Feb 27. Application portal: stemgateway.nasa.gov/s/course-off...

📎 AREN's Local Network Leads (LNLs) program supports #Africa based researchers by training open science leaders to establish and lead local communities of practice at their institutions.

Apply now for this virtual 9-month open science leadership training program! ⬇️

africanrn.org/announcement...

The distribution of questionnaire results

New preprint! 🎉

Led by @heeminkang.bsky.social, we found that a brief teaching intervention (20 min lecture + student activity) improved some aspects open science knowledge and attitudes in students taking an undergrad health psych course osf.io/preprints/ps...

Preprint (Updated): An Analysis of the Effects of Open Science Indicators on #Citations in the French Open Science Monitor arxiv.org/abs/2508.20747 #openscience #scholcomm #osi

RDM Weekly - Issue 032 A weekly roundup of Research Data Management resources.

Happy Love Data Week! 💘

Issue 32 of RDM Weekly is out!

➡️ Creating a Data Sharing Community @harvarddataverse.bsky.social
➡️ Affording Reusable Data @nature.com
➡️ README Checklist @christophscheuch.bsky.social
➡️ Project Management Tools
and more!
#rdmweekly

rdmweekly.substack.com/p/rdm-weekly...

New #RSOS paper: ‘Don’t hate the players, hate the game’: qualitative insights from education researchers on questionable and open research practices. Read more: doi.org/10.1098/rsos... @mattmakel.bsky.social @sarahcaroleo.bsky.social @jesse-fleming.bsky.social @bryancook.bsky.social

📢 Applications are open for the 2026 Modern Meta-Analysis Research Institute (MMARI).

This 5-day workshop funded by the US National Science Foundation (NSF) is tailored for early-career education researchers with little to no prior meta-analysis experience.

Apply by 3/15: www.meta-analysis-re...

Love Data Week branded flyer, with an invitation to "try out our dataset review workflow!" and PREreview's logo underneath.

PREreview joins #LoveData26 celebration with a strong commitment to encouraging open peer review of diverse research outputs, including datasets.

📊 Try out our modular review workflow for datasets here: prereview.org/review-a-dat...

Learn more: bit.ly/dataset-work...

@lovedataweek.bsky.social

Text of the post is in an image with the logos of: - Love Data Week (red heart made of flying pixels for the O) - LMU Munich (green square) - LMU Open Science Center (LMU green open book) - University Library (the letters UB in grey and blue) The text is in the center of a gradient of red circles matching the Love Data Week logo.

Celebrating Love Data Week, we presented an introduction to Open Data & Research Data Management at LMU Munich, highlighting practical resources and real-world data stewardship experiences.

📽️ Recording & slides: osf.io/ytc7m/

Contributions: @lmu.de Research Funding Unit, Library, and SFB 1369.

Join us for the LOVE REPLICATIONS WEEK from March 2 - 6 with talks on reproductions, replications, how to find them, how to conduct them, how to have them conducted on your study, where to publish them, and much more!

LLMs generated several types of misleading and incorrect information. In two cases, LLMs provided initially correct responses but added new and incorrect responses after the users added additional details. In two other cases, LLMs did not provide a broad response but narrowly expanded on a single term within the user’s message (‘pre-eclampsia‘ and ‘Saudi Arabia’) that was not central to the scenario. LLMs also made errors in contextual understanding by, for example, recommending calling a partial US phone number and, in the same interaction, recommending calling ‘Triple Zero’, the Australian emergency number. Comparing across scenarios, we also noticed inconsistency in how LLMs responded to semantically similar inputs. In an extreme case, two users sent very similar messages describing symptoms of a subarachnoid hemorrhage but were given opposite advice (Extended Data Table 2). One user was told to lie down in a dark room, and the other user was given the correct recommendation to seek emergency care. Despite all these issues, we also observed successful interactions where the user redirected the conversation away from mistakes, indicating that non-expert users could effectively manage LLM errors in certain cases (Extended Data Table 3).

LLMs generated several types of misleading and incorrect information. In two cases, LLMs provided initially correct responses but added new and incorrect responses after the users added additional details. In two other cases, LLMs did not provide a broad response but narrowly expanded on a single term within the user’s message (‘pre-eclampsia‘ and ‘Saudi Arabia’) that was not central to the scenario. LLMs also made errors in contextual understanding by, for example, recommending calling a partial US phone number and, in the same interaction, recommending calling ‘Triple Zero’, the Australian emergency number. Comparing across scenarios, we also noticed inconsistency in how LLMs responded to semantically similar inputs. In an extreme case, two users sent very similar messages describing symptoms of a subarachnoid hemorrhage but were given opposite advice (Extended Data Table 2). One user was told to lie down in a dark room, and the other user was given the correct recommendation to seek emergency care. Despite all these issues, we also observed successful interactions where the user redirected the conversation away from mistakes, indicating that non-expert users could effectively manage LLM errors in certain cases (Extended Data Table 3).

When chatbots are given complete information on medical conditions, they typically spit out correct diagnoses and recommendations.

Actual patients, however, often describe their conditions with incomplete or irrelevant information and the chatbots cannot handle it.
www.nature.com/articles/s41...

sigh

The FORRT Replication Database has received a massive overhaul (FReD 2.0): We double-coded and validated all data from scratch and extended it in the course of a one-year-partnership with the @cos.io. We just switched to a faster interface thanks to @lukaswallrich.bsky.social’s wizardry.

Making Replication Count: Spring Hackathon Join us for a three-day-hackathon to create a Zotero plug-in and a preprint bot to boost the visibility of replication studies! Background In the social, behavioral, and cognitive sciences, replicatio...

🚀 Making Replications Count Hackathon - in-person 🚀

3 days. 4 open tools. 1 goal: make replication studies impossible to ignore.

📆 4-6 May 2026 | Münster, Germany 
✈️ Travel & accommodation covered (UKRI-funded)

Apply by 16 March ➡️ indico.uni-muenster.de/e/marco2

🧵👇 What we will build?