Search: “coppa”

34 stories

DomainAll Cyber AIWindow24h 3d 7d 30d AllIn the wild Important only Show duplicates

CategoryAll Policy & legal 70 Threat actor 8 Data breach 7 Phishing & fraud 3 Industry 3 Ransomware 3 Research 2 Vulnerability 2 Malware 2 Exploit / PoC 1 Advisory 1

SeverityAll critical 17 high 25 medium 3 low 54 info 3

Smart search ranks by meaning as well as keywords (one row per story, last 45 days).

TikTok Settles U.S. Child Privacy Case for $400 Million

TikTok will pay $400 million to settle U.S. DOJ/FTC claims that it violated COPPA by collecting data from children under 13.

The U.S. Department of Justice announced a $400 million settlement with TikTok and ByteDance resolving a 2024 lawsuit over violations of the Children's Online Privacy Protection Act (COPPA). TikTok will pay $300 million immediately and $100 million upon entry of an order vacating a prior consent decree against its predecessor Musical.ly; it is one of the largest recoveries ever obtained in a COPPA case. The DOJ and FTC, filing in California, alleged TikTok knowingly allowed children under 13 to create accounts and illegally collected data via Kids Mode. TikTok was previously fined €345 million by Ireland's Data Protection Commission in 2023 for GDPR breaches involving children's data.

Security AffairsAug 24, 07:23 UTC · 24d agoPolicy & legal

PaperCut NG/MF Critical Zero-Day Exploited in the Wild

PaperCut confirms active exploitation of critical zero-days CVE-2026-81578 and CVE-2026-82078 in PaperCut NG/MF print management software, with confirmed customer incidents.

PaperCut Software issued an urgent advisory on August 27, 2026, confirming it is investigating active exploitation of flaws in PaperCut NG and PaperCut MF and treating the issue as a security emergency. On August 28 the vendor assigned CVE-2026-81578 and CVE-2026-82078 to the two vulnerabilities forming the exploit chain. At initial disclosure, no CVSS score, vulnerability class, authentication requirements, or exploit path details were public. Rapid7 is tracking the issue as an emerging threat.

Rapid7 BlogAug 28, 10:09 UTC · 20d agoExploit / PoC in the wildCVE-2026-81578 CVE-2026-82078

TikTok Agrees to $400 Million Settlement in U.S. Child Privacy Lawsuit

TikTok agrees to pay $400 million to settle a DOJ child-privacy lawsuit alleging COPPA violations including collecting data from children under 13.

The DOJ announced that ByteDance-owned TikTok will pay $400 million to settle a 2024 lawsuit alleging massive-scale invasions of children's privacy. TikTok pays $300 million immediately and $100 million upon vacating the prior Musical.ly consent decree. The complaint alleged children under 13 could create accounts, data was collected in Kids Mode, and parental deletion requests went unheeded. DOJ called it one of the largest recoveries ever under COPPA, following TikTok's 2023 €345 million GDPR fine.

The Hacker NewsAug 22, 14:32 UTC · 25d agoPolicy & legal

Forgery of C2PA on a Pixel 10

Researcher forged a Google Pixel 10 C2PA content credential with genuine signatures, showing root-level attackers can fake photo provenance.

A Hacker Factor blog post demonstrates an AI-generated 'unicorn glitter milk' news photo carrying a valid, cryptographically signed C2PA manifest traceable to Google's Pixel camera certificate chain, passing validation in Adobe Inspect and the CAI Verify tool with a verified timestamp. The author, working with UMBC's PASAWG working group, reported to Google and C2PA in November 2025 that root access on a Pixel device could sign arbitrary images as camera captures; after 90 days without resolution, details were published. The finding undermines C2PA Assurance Level 2 claims made for Pixel 10 Content Credentials.

Lobsters · securitySep 16, 13:24 UTC · 1d agoResearch

Meta pledges to overhaul kids’ safety protections, pay $17 billion to settle social media case

Meta settles states' kids' online safety lawsuit for $17 billion, agreeing to landmark usage limits, age restrictions, and independent auditing.

Meta agreed to pay $17 billion to settle a civil suit from nearly every US state and territory alleging it hid research showing Facebook and Instagram are addictive to minors and violated COPPA by collecting data on children under 13. The settlement imposes reforms including two-hour daily limits for users under 18, a midnight-to-6am usage block, non-personalized feed options, and an independent auditor. Meta also settled separately with Texas for about $1 billion.

The RecordAug 26, 18:53 UTC · 21d agoPolicy & legal

PaperCut Zero-Day: Active Exploitation and Pre-Auth RCE

PaperCut NG/MF hit by a pre-auth RCE zero-day under active exploitation; Huntress reproduced the chain and urged immediate patching.

Huntress reports active exploitation of a zero-day in PaperCut NG and PaperCut MF, and says it reproduced a pre-authentication remote code execution chain. The flaw allows unauthenticated attackers to execute code on exposed PaperCut servers. Huntress published urgent patching, exposure-reduction, and detection guidance. No CVE identifier was provided in the announcement.

HuntressAug 28, 03:00 UTC · 20d agoExploit / PoC in the wild

Risky Bulletin: Expired cards can be used for new transactions

Researchers show expired Visa contactless cards can be revived via NFC man-in-the-middle relay to run fraudulent transactions; roundup also covers major breaches.

University of Massachusetts Amherst researchers built an NFC man-in-the-middle rig that updates a card's expiration date in transit and relays the modified payment to POS terminals, reviving expired contactless cards; Visa terminals and the backends of all five banks studied failed to catch the manipulation. The same roundup reports Iranian hackers shut down a small UK power plant for four days, Lazarus breached South Korea's Presidential Office as part of a campaign exceeding 100 victims, and French telecom SFR suffered a breach affecting over 2.1 million customers.

Risky Business NewsAug 24, 04:40 UTC · 24d agoResearch1

U.S. CISA adds PaperCut NG/MF flaws to its Known Exploited Vulnerabilities catalog

CISA added two actively exploited PaperCut NG/MF pre-auth flaws to the KEV catalog; federal agencies must patch by September 14.

CISA added CVE-2026-81578 (CVSS 8.8, missing authentication for critical function) and CVE-2026-82078 (CVSS 9.4, unsafe reflection) in PaperCut NG/MF to its Known Exploited Vulnerabilities catalog. Huntress confirmed active pre-authentication RCE exploitation in two customer environments and reproduced the full chain against a clean PaperCut NG 25.0.11 server, chaining the auth bypass into unsafe Java class loading for SYSTEM-level execution. About 47% of roughly 2,500 tracked PaperCut installs still run version 23 or earlier with no patch available, and observed attacker activity was limited to system discovery commands.

Security AffairsSep 1, 08:13 UTC · 16d agoVulnerability in the wildCVE-2026-81578 CVE-2026-82078

RoboSPA: Can VLA Models Go Beyond Simple Scenes and Short-Horizon Tasks?

RoboSPA introduces a 527K-trajectory benchmark with 280 task variants showing current VLA models struggle with spatial reasoning and long-horizon planning.

RoboSPA is a large-scale robotic manipulation dataset and benchmark for diagnosing embodied reasoning in vision-language-action models, covering fine-grained spatial reasoning and long-horizon procedural planning. It spans 10 task categories and 56 base tasks instantiated across five difficulty levels for 280 variants, with 527K trajectories collected across multiple embodiments and diverse scenes. Diagnostic metrics go beyond binary success rates, and experiments show current VLA models struggle with complex spatial relations, precise low-level execution, and memory-intensive planning.

Hugging Face daily papersSep 3, 20:00 UTC · 13d agoAI research

CVE-2026-81578 + CVE-2026-82078 | PaperCut NG/MF Authentication Bypass and Unsafe Dynamic Class Loading Vulnerabilities

Two chained PaperCut NG/MF flaws, CVE-2026-81578 and CVE-2026-82078, enable pre-auth RCE, with active exploitation confirmed in customer environments.

PaperCut NG/MF is affected by CVE-2026-81578, an improper access control flaw (CVSS 4.0 8.8) allowing unauthenticated configuration changes, and CVE-2026-82078, an unsafe dynamic class loading flaw (CVSS 4.0 9.4) enabling arbitrary Java bytecode execution. Chained, they yield pre-authentication remote code execution on the PaperCut Application Server. PaperCut confirmed active exploitation and customer incidents, and Huntress observed exploitation starting August 26, 2026. Emergency Patch Release 2 is available for NG/MF v24-v26, with IOCs including suspicious pc-app.exe child processes, truncated server.log files, and AnyDesk installs.

Horizon3.aiSep 1, 15:31 UTC · 15d agoExploit / PoC in the wildCVE-2026-81578 CVE-2026-820781

VU#456290: Hugging Face Transformers library writes remote code to disk prior to consent check

CVE-2026-80047: Hugging Face Transformers 4.49.0 through 5.8.1 writes attacker-controlled Python files to disk before the trust_remote_code consent check.

CERT/CC vulnerability note VU#456290 describes CVE-2026-80047 in the Hugging Face Transformers library, affecting versions 4.49.0 through 5.8.1. The library performs a remote module fetch and writes attacker-controlled Python files to the local disk before evaluating the trust_remote_code consent prompt, without user authorization. This violates the security contract enforced across other dynamic module-loading paths in the library. Transformers is a primary framework for training and inference across NLP, vision, audio, video, and multimodal machine learning systems.

CERT/CC Vulnerability NotesSep 1, 13:35 UTC · 16d agoVulnerability CVE-2026-800471

Multi-Vector (Late Interaction) Embedding Models with Sentence Transformers

Hugging Face details building and using multi-vector late-interaction embedding models with Sentence Transformers for retrieval workloads.

Hugging Face published a guide on multi-vector, late-interaction embedding models (ColBERT-style) supported through Sentence Transformers. The post covers how practitioners can build and use these models for retrieval and RAG pipelines. It is a developer tooling and technique write-up, not a security advisory.

Hugging Face BlogAug 18, 00:00 UTC · Aug 18, 2026AI tools & infra1

PaperCut NG/MF vulnerabilities exploited in zero-day attacks

PaperCut warns of active zero-day exploitation chaining CVE-2026-81578 and CVE-2026-82078 for pre-auth remote code execution in NG/MF print management.

PaperCut Software confirmed attackers are chaining two vulnerabilities in PaperCut NG and MF: CVE-2026-81578, an improper access control flaw in the web management interface allowing unauthenticated configuration changes, and CVE-2026-82078, unsafe dynamic class loading in database connection utilities enabling arbitrary Java bytecode execution. Huntress reproduced a pre-authentication remote configuration takeover and full RCE chain against PaperCut NG 25.0.11.75758 and observed limited exploitation at two customers, including post-exploitation whoami and ver commands. The vendor released Emergency Patch Release 2 with additional hardening and urged restricting Application Server web access to trusted IPs. In 2023, Clop and LockBit affiliates abused CVE-2023-27350 and CVE-2023-27351 in the same software.

Help Net SecurityAug 31, 14:39 UTC · 16d agoExploit / PoC in the wildCVE-2026-81578 CVE-2026-82078 CVE-2023-27350+1 CVEs1

PaperCut NG/MF Zero-day Vulnerability Exploited in the Attacks (CVE-2026-82078 & CVE-2026-81578)

PaperCut NG/MF zero-days CVE-2026-82078 and CVE-2026-81578 are actively exploited, chaining to pre-auth RCE on the Application Server.

PaperCut issued an urgent security advisory for two actively exploited vulnerabilities, CVE-2026-82078 and CVE-2026-81578, affecting all versions of PaperCut NG and PaperCut MF. When chained, the flaws enable pre-authentication remote code execution on the PaperCut Application Server, originating from a request-confusion and authorization gap. PaperCut is widely deployed for print management in enterprises and education, making unpatched instances a high-value target.

Qualys ThreatPROTECTAug 31, 12:07 UTC · 17d agoExploit / PoC in the wildCVE-2026-82078 CVE-2026-81578

CISA Adds Two Known Exploited Vulnerabilities to Catalog

CISA added actively exploited PaperCut NG/MF flaws CVE-2026-81578 and CVE-2026-82078 to the KEV catalog, mandating federal patching.

CISA added two vulnerabilities to its Known Exploited Vulnerabilities catalog based on evidence of active exploitation: CVE-2026-81578 (PaperCut NG/MF missing authentication for critical function) and CVE-2026-82078 (PaperCut NG/MF unsafe reflection). Under Binding Operational Directive 26-04, Federal Civilian Executive Branch agencies are required to prioritize and apply these updates. The KEV listing signals observed exploitation of the PaperCut print management platform.

CISA AdvisoriesAug 31, 12:00 UTC · 17d agoExploit / PoC in the wildCVE-2026-81578 CVE-2026-82078

Hackers Are Probing PaperCut Servers, and 47% Still Have No Patch

PaperCut pre-auth RCE CVE-2026-81578 actively exploited; 47% of tracked installs unpatched and exposed to remote code execution.

Huntress confirmed active exploitation of a pre-authentication RCE flaw in PaperCut (CVE-2026-81578), chainable with CVE-2026-82078 to achieve SYSTEM-level code execution via PaperCut's Application Server. Observed attacks focused on short reconnaissance (whoami, ver, tasklist) with no persistence, but a malicious Java class dropping into the installation directory leaves the jdbc:derby:memory:pwn breadcrumb. Roughly 47% of ~2,500 tracked installations run version 23 or earlier with no patch yet available.

Security AffairsSep 6, 08:17 UTC · 11d agoExploit / PoC in the wildCVE-2026-81578 CVE-2026-82078

Adobe security advisory (AV26-848)

Canada's Cyber Centre relayed Adobe advisories covering vulnerabilities in Campaign Classic, Substance 3D apps, Adobe XD, Illustrator, and C2PA tools.

Bulletin AV26-848 lists Adobe vulnerabilities affecting Campaign Classic (through 7.4.4 build 9400), Substance 3D Designer, Painter, and Sampler, Adobe XD, C2PA Tool, Content Credentials Rust SDK, and Illustrator 2025/2026. The Canadian Centre for Cyber Security encourages users and administrators to review the linked Adobe bulletins and apply updates. Specific CVE identifiers are not enumerated in the advisory text.

Canadian Centre for Cyber SecurityAug 26, 13:56 UTC · 22d agoAdvisory

RoboSPA: Can VLA Models Go Beyond Simple Scenes and Short-Horizon Tasks?

RoboSPA introduces a 527K-trajectory robotic benchmark with 280 task variants exposing VLA model weaknesses in spatial reasoning and long-horizon planning.

RoboSPA is a large-scale robotic manipulation dataset and benchmark for diagnosing embodied reasoning in vision-language-action models, spanning fine-grained spatial reasoning and long-horizon procedural planning. It covers 10 task categories and 56 base tasks instantiated across five difficulty levels for 280 variants, with 527K trajectories collected across multiple embodiments. Experiments on representative VLA models show current systems struggle with complex spatial relations, precise low-level execution, and memory-intensive planning.

arXiv cs.AI / cs.LG / cs.CLSep 4, 16:19 UTC · 12d agoAI research

LexFlip: A Dissociation Diagnostic for Legal Meaning Preservation Metrics

LexFlip releases 373 minimal perturbations of Quebec statutory French that reverse legal force while preserving tokens, exposing weaknesses in embedding-based meaning preservation metrics.

LexFlip provides 373 minimal perturbations of Quebec statutory French that reverse legal force while preserving 0.93 of tokens, creating dissociation items that break monotone token-overlap metric validation. The seven embedding and BERTScore metrics tested register only 0.022-0.039 of their identical-to-unrelated range on these edits, versus 0.670 for bidirectional NLI. Against FrJudge, with a measured human ceiling of r=0.597, a bare length feature outscores every semantic metric tested.

arXiv cs.AI / cs.LG / cs.CLSep 4, 15:49 UTC · 12d agoAI research

libpcap 1.10.7 fixes 7 vulnerabilities

libpcap 1.10.7 fixes seven vulnerabilities, including CVE-2026-31912 affecting the deprecated bpf_filter(), and all users are advised to update.

libpcap 1.10.7 was published on 2026-09-05 with its main focus being fixes for seven vulnerabilities. Maintainer Denis Ovsienko announced the release on the oss-security mailing list, noting that each CVE record contains detailed accounts and links to git commits with further explanation. For the CVE-2026-31912 fix to take effect, applications using the now deprecated bpf_filter() must update their usage. All users are advised to update; no exploitation in the wild is reported.

oss-securitySep 9, 01:35 UTC · 8d agoVulnerability CVE-2026-319122

PaperCut security advisory (AV26-858) – Update 2

CISA added actively exploited PaperCut MF/NG flaws CVE-2026-81578 and CVE-2026-82078 to KEV; Canada's Cyber Centre urges Emergency Patch Release 2 for v24-v26.

Canada's Cyber Centre updated security advisory AV26-858, noting PaperCut MF and PaperCut NG versions prior to the v24, v25, and v26 Emergency Patch Release 2 are affected by CVE-2026-81578 and CVE-2026-82078. Open-source reporting indicates both vulnerabilities are being exploited in the wild. In Update 2 (August 31, 2026), CISA added both CVEs to its Known Exploited Vulnerabilities catalog, and the Cyber Centre urges administrators to apply the emergency patches immediately.

Canadian Centre for Cyber SecurityAug 31, 16:00 UTC · 16d agoExploit / PoC in the wildCVE-2026-81578 CVE-2026-82078

PaperCut Zero-Day Under Active Attack: Emergency Patch Released

PaperCut warns that a zero-day in NG and MF print management products is being actively exploited, shipping emergency patches with no CVE assigned yet.

PaperCut Software confirmed active exploitation of a zero-day affecting PaperCut NG and MF, with confirmed customer incidents, and released emergency patches on Friday ahead of technical details or a CVE assignment. The company urges customers to restrict web access to trusted IP addresses, disconnect application servers from the internet, and watch for indicators including tampered or truncated server.log files, database error strings, and suspicious activity involving pc-app.exe. The article recalls the May 2023 exploitation of CVE-2023-27350 and CVE-2023-27351 by the financially motivated Lace Tempest group, a Clop ransomware RaaS affiliate, though no attacker attribution has been given for the current campaign.

Security AffairsAug 29, 20:42 UTC · 18d agoExploit / PoC in the wildCVE-2023-27350 CVE-2023-273511

How Hugging Face Inference Endpoints, Jobs, and Buckets Power Search on Papers with Code

Hugging Face explains how Inference Endpoints, Jobs, and Buckets power semantic search on Papers with Code.

Hugging Face describes the infrastructure behind search on Papers with Code, built on its Inference Endpoints, Jobs, and Buckets services. The post is a product-focused engineering walkthrough with no security impact.

Hugging Face BlogAug 21, 00:00 UTC · 27d agoAI tools & infra

Researchers Use Claude to Port Pre-Auth RCE Exploit From One PLC Model to Another

Forescout used Anthropic's Claude to port a working pre-auth RCE exploit for CVE-2021-31886 between WAGO PLCs, executing ARM shellcode on live hardware.

Forescout Vedere Labs used Claude interactively to port an RCE exploit for CVE-2021-31886, a CVSS 9.8 stack buffer overflow in the Nucleus FTP server's USER command, from a WAGO 750-852 to a WAGO 750-831 PLC, running attacker-supplied ARM shellcode. The final RCE stage cost $535.74 in API usage over 8 hours 32 minutes, and a follow-up attempt to build a C2 implant permanently bricked the device. CERT@VDE lists many Nucleus V1-based WAGO models as vulnerable with no updates available; Siemens plans no Nucleus NET remediation. The work follows a joint NSA/CISA/FBI/DOE/EPA advisory warning of AI-generated exploitation scripts targeting internet-exposed Siemens S7 PLCs.

The Hacker NewsSep 2, 07:47 UTC · 15d agoResearch CVE-2021-318862

Fake CVE-2023

A fake PoC for WinRAR RCE CVE-2023-40477 posted on GitHub actually deploys VenomRAT through a multi-step infection chain targeting researchers.

Four days after Zero Day Initiative publicly disclosed the WinRAR RCE vulnerability CVE-2023-40477 on August 17, 2023, an actor using the alias whalersplonk published a fake proof-of-concept on GitHub. The Python script actually repurposed public PoC code for GeoServer SQL injection CVE-2023-25157 and triggered an infection chain ending in VenomRAT. The README and an accompanying video lured users into running the script; the video drew over 100 plays. Unit 42 assesses the actor was opportunistic, targeting other miscreants adopting new vulnerabilities rather than researchers specifically.

Palo Alto Unit 42Aug 17, 10:57 UTC · Aug 17, 2026Malware CVE-2023-40477 CVE-2023-25157

Legora reviewed 41 documents in minutes with GPT-6 Astra

Legal-tech firm Legora says GPT-6 Astra reviewed 41 financial documents in minutes, catching all four planted errors and boosting accuracy about 40%.

Legal technology company Legora reported using OpenAI's GPT-6 Astra to review 41 financial-statement documents in minutes. The workflow found all four planted errors and improved performance by nearly 40% compared to prior processes. The case study highlights AI-assisted financial review adoption in professional services.

OpenAI NewsSep 3, 12:00 UTC · 14d agoAI industry

The OCUDU dApp Platform: An Open Runtime and E3 Interface for Real-Time AI-RAN

OCUDU open runtime lets third-party signed AI-RAN dApps run inside production 5G distributed units under three timing contracts, released as BSD-3 preview.

The OCUDU dApp platform provides an open runtime and E3 interface for executing signed AI-RAN applications inside a production 3GPP NR distributed unit, where prior dApp frameworks could only observe export streams. Modules run under three typed timing contracts: GPU receive-chain residency (Class A), the scheduler's 100 microsecond deadline (Class B), or non-blocking observer (Class C). On a GB10 gNB, dApps including an out-of-tree neural equalizer ran on a live cell without fallback. Platform, SDK, and a zero-hardware quickstart are public under BSD-3-Clause-Clear as a preview of the OCUDU AI-RAN Working Group 2.

arXiv cs.AI / cs.LG / cs.CLSep 7, 18:03 UTC · 9d agoAI tools & infra1

In-Context Robot Learning with VLM Agents

GPT-Policy uses commercial VLM agents like GPT-6 Astra for in-context robot learning, translating demonstrations and feedback into verified robot actions without gradient updates.

The paper (arXiv 2609.19138) presents GPT-Policy, a framework combining a context compiler that preserves task-relevant visual transitions, a VLM such as GPT-6 Astra that proposes robot-tool actions, and a constrained controller that verifies and executes each action. Real-robot trials show human video demonstrations improve task completion even without robot action labels, while aligned action references yield further gains on contact-sensitive tasks. Evaluation covers success and efficiency metrics, matched model comparisons, and controlled context ablations.

Hugging Face daily papersSep 15, 20:00 UTC · 1d agoAI research

Right Tool, Right Job: Native-Language Evaluation, Tokenizer Sensitivity, and Methodological Findings from a French-Only BabyLM

French BabyLM entry METRON-FR (125M GPT-2, 92.47M words) shows tokenizer artifacts dominate child-scale zero-shot evaluation; proposes standard diagnostics.

METRON-FR is a 125M-parameter GPT-2 pretrained on 92.47M French words, submitted to the BabyLM 2026 Strict track, scoring 85.97% on the native Quebec-French QFrBLiMP benchmark and 62.80% on the BabyLM-weighted leaderboard. A cross-lingual GLUE protocol combining French task-data translation with rank-16 LoRA shows relational tasks gain while world-knowledge tasks regress. Bilingual Lexicon Induction reaches p@1 of 68.84%, 18x above chance, and ablations show single-token zero-shot scoring is dominated by tokenizer and template artifacts at child scale.

arXiv cs.AI / cs.LG / cs.CLSep 15, 16:43 UTC · 1d agoAI research

Re: pcre2 version 10.48 released with security fixes

PCRE2 10.48 released with security fixes; seven vulnerabilities assigned CVEs including CVE-2026-86145 and CVE-2026-89156 through CVE-2026-89162.

The PCRE2 regular expression library released version 10.48 with multiple security fixes, disclosed on the oss-security mailing list. Seven issues received CVE assignments: CVE-2026-86145, CVE-2026-89156, CVE-2026-89157, CVE-2026-89158, CVE-2026-89160, CVE-2026-89161, and CVE-2026-89162. PCRE2 is a widely used regex engine embedded in many open-source projects and applications. No exploitation is reported in the disclosure.

oss-securitySep 11, 20:19 UTC · 5d agoVulnerability CVE-2026-86145 CVE-2026-89156 CVE-2026-89157+4 CVEs1

Structured Claim-Level Discourse Representations for Dense Health Narratives

Researchers propose a claim-level discourse framework for health videos, finding 13.22 atomic claims per minute and that LLMs struggle with pragmatic profiling.

The paper introduces a structured framework for claim-level discourse analysis in dense health narratives on social media videos, modeling tuples that link atomic claims with thematic aspects, stance, and multidimensional pragmatic attributes. Analysis found an average of 13.22 atomic claims per minute in health video discourse. A benchmark spanning four health domains with 1,191 manually annotated claims from 60 videos shows current LLMs perform strongly on thematic categorization and stance prediction but struggle with high-dimensional pragmatic profiling, suggesting future systems need task decomposition and specialized inference strategies.

arXiv cs.AI / cs.LG / cs.CLSep 16, 16:43 UTC · 21h agoAI research

Can Foundation Models Moderate Online Content? Evaluating Instruction- vs. Example-Driven Policy Operationalization

ModerationBench shows foundation models can nearly triple Bluesky's moderation F1 (0.60 vs 0.22), with instruction- and example-driven guidance performing comparably.

Researchers built ModerationBench, a new benchmark of 4,000 manually annotated in-the-wild posts from Bluesky, to test whether foundation models can reliably operationalize content moderation policies. They systematically compare instruction-driven guidance (reasoning from policy precepts) with example-driven guidance (generalizing from precedents) for Vision-Language Models. Both paradigms achieve comparable peak effectiveness, and foundation models nearly triple the F1 of Bluesky's deployed moderation system on Random Posts (0.60 vs 0.22).

arXiv cs.AI / cs.LG / cs.CLSep 9, 16:30 UTC · 7d agoAI research1

Performance of Clinical AI System and Physicians and Frontier Language Models in primary care diagnostics

Clinical AI system Doctorina achieved 82.0% primary-care diagnostic concordance versus 57.0% for physicians across 150 synthetic consultations.

The study compared Doctorina, eight physicians, and four standalone frontier language models on 150 synthetic Polish-language primary-care consultations. Doctorina achieved 82.0% Top-1 diagnostic concordance versus 57.0% for physicians (25.0-point difference, 95% CI 17.7-32.7) and 97.3% versus 85.0% primary-or-reference-differential concordance. Normalized workup and treatment scores were 89.4 versus 66.9 and 83.7 versus 61.2. Kimi K3 ranked next on diagnosis, while Claude Opus 5 led the closely spaced management estimates among Opus, Doctorina and Kimi.

arXiv cs.AI / cs.LG / cs.CLSep 8, 17:25 UTC · 8d agoAI research

Semigroup-JEPA: Latent Dynamics Consistency for Zero-Shot Physics Generalization

SG-JEPA world model conditions latent prediction on physical parameters, halving open-loop prediction error versus DINO-WM and boosting robotic control success.

SG-JEPA extends the LeWorldModel JEPA framework by supplying the governing physics parameter to the temporal model via action-conditioning and jointly training an encoder and predictor through autoregressive latent rollout. On out-of-distribution gravitational-field tasks it reduces open-loop prediction error by up to 2x versus DINO-WM on 2D datasets and increases 3D robotic control success rate up to 2.5x using independently trained diffusion policies. A linear feature analysis attributes most of the gain to the encoder learning features that the predictor can carry forward through rollout.

arXiv cs.AI / cs.LG / cs.CLSep 9, 17:08 UTC · 7d agoAI research