AI News Archive: May 24, 2026 — Part 6
Sourced from 500+ daily AI sources, scored by relevance.
- Almo AI
Your personal AI that searches the web and learns
- Calsera AI
calorie,calsera,calsera ai, calorie tracker,
- Clever AI
Undetectable AI copilot for interviews & meetings
- DocuGen AI
AI that writes your code documentation instantly
- Saidrix AI Tutor
Your Personal AI-Powered Learning Companion
- Reality Firewall by Mosam Tech
Spot AI scams before they cost you money.
- Grillr.io
The AI accountability partner that hunts you down.
- DescriptoAI - Descriptions produits IA
Générez des descriptions SEO en 10 secondes
- Nova by Altura
The AI that actually executes.
- WealthOS Elite
WealthOS Elite — Your AI-Powered Wealth Command Center
- podsML
one-click deployment for everything AI
- Hemlo
most accurate prediction of future by simulating an ai world
- Health AI Insight
"Health AI Insight" - bridges Apple HealthKit and AI.
- harnessforge
Generate harness files for AI coding agents
- VISNIB
See how ChatGPT, Perplexity & AI Overview rank your store
- AIDLE by LOBSTERR
Daily word puzzle for AI developers & Web3 builders
- OpenLoad
OpenLoad — Docs, Sheets, Slides, PDFs, all powered by AI.
- FreightLoop
AI freight marketplace for shippers and carriers
- BeatMind
The fully offline AI music studio for macOS
- textools SEO Audit
Weekly-updated 24-point SEO audit with AI Visibility analysi
- CoinCup
Predict World Cup 2026 matches & compete with friends
- HOW
Your Tiny AI that can use your deskop .
- LoraDB
Embedded Rust graph database for AI, Cypher & Vectors Search
- SayCopyPaste
Your voice, instantly pasted.
- Token Relay
One API Key, All AI Models. Pay with Crypto.
- Nova AI Companion
Privacy-First, Portable AI-Companion with Persistent Memory
- Moodifyr
The Music Player That Understands Your Mood
- Gym control
AI-powered gym management made simple and smart
- Will AI Replace Radiology Professionals?
ai vs medical imaging technologist
- CCMNet
CCMNET – Free Online Tools
- RepoLens
RepoLens helps developers understand GitHub repository.
- vid2vid
Video to video, zero threshold for creativity
- WatermarkRemover.live
Remove watermarks from images & videos free — AI-powered
- Discovery Music
Incubateur IA pour artistes indépendants
- MCPSpend
Know what your AI agents really cost
- 织界 WorldLoom
An AI-powered multi-world narrative engine
- Avmira
AI Learning, Developer Tools & Verifiable Certificates
- UXBud
AI-Powered UX Intelligence
- aicvmakers.in
AI ATS Resume Builder India | Resume for Naukri & MNC Jobs
- Wonder Alarm
Cinematic customization and intelligent sleep app
- logsgen
LogSgen AI is the first AI operating system built for global
- KingSSP
Free Tools, AI Tools, Free Tools Online
- Video to Prompt
Break down videos into individual scenes convert to prompt.
- PixelMax
Edit Your Photos With AI.
- Vtuber Assistant
Your AI VTuber co-host that actually pays attention.
- Genos-m: a foundation model for human-associated microbial genomes
Human-associated microbial genomes encode extensive strain-level diversity and niche-specific gene repertoires that are critical to host health. However, these complex sequence features remain difficult to capture using general-purpose DNA foundation models, highlighting the need for dedicated representation learning tailored to the human microbiome. Here, we introduce Genos-m, an open-source foundation model for human-associated microbial genome representation. Genos-m was pretrained on approximately 1.2 trillion nucleotide tokens from a curated microbial genome corpus, including human-associated prokaryotic isolates, high-quality metagenome-assembled genomes (MAGs) and bacteriophages, supplemented with GTDB species-level representative genomes to broaden prokaryotic taxonomic breadth. The model uses a sparsely activated Mixture-of-Experts (MoE) Transformer architecture, with 4.7 billion total parameters, approximately 330 million activated parameters per forward pass and a maximum context length of one million base pairs. We evaluated frozen Genos-m representations across short-sequence and gene-level tasks, biosynthetic gene cluster (BGC)-based regional sequence tasks, whole-genome strain phenotype prediction, and zero-shot transfer on prokaryote-related RNAfitness assays. Across these benchmarks, Genos-m consistently ranked among the leading comparison models, with the best performance in five of eight gene-fitness regression tasks and in BGC type classification. Using sparse autoencoders, we identified sparse features in Genos-m hidden activations that aligned with annotated ORFs, intergenic regions, and tRNA and rRNA loci. In downstream applications, Genos-m-derived genome-informed species representations incorporated into a human microbiome self-supervised learning model improved colorectal cancer (CRC)-control classification over conventional species-abundance random forest models. Genos-m also generated stable sample-level embeddings from as few as 10,000 metagenomic reads, retaining gut microbial community structure that distinguished geographic origin and aligned with enterotypes defined from full-depth taxonomic profiles. Together, these results support Genos-m as a reusable representation model for microbial genomes and metagenomes, with conclusions bounded by the reported datasets, task definitions and evaluation protocols. Genos-m model weights, inference code, and usage documentation are publicly available on GitHub (https://github.com/BGI-HangzhouAI/Genos-m) and HuggingFace (https://huggingface.co/BGI-HangzhouAI/Genos-m).
- Spatio-temporal machine learning for multi-horizon prediction of bluetongue outbreaks
Reliable early warning of infectious disease outbreaks remains a major challenge for surveillance systems, particularly for vector-borne pathogens whose transmission depends on interactions among hosts, vectors, and climate-sensitive environmental conditions. Data-driven forecasting offers a promising approach for predicting outbreak risk using surveillance and environmental data. This study develops a logit-weighted ensemble (LWE), a machine-learning framework that predicts outbreak occurrence 1 -- 6 months ahead at the administrative unit -- month scale using routinely available outbreak notifications and gridded climate data. Bluetongue virus (BTV), an arbovirus of ruminants transmitted by Culicoides biting midges, provides a well-characterised system in which transmission is strongly shaped by climate, making it a useful system for applying and testing this approach. The framework is evaluated using surveillance data collected between 2005 and 2024 from France, Greece, and Italy, selected for their long-running and high-quality outbreak surveillance records. Across all three countries, the LWE achieved the strongest and most stable predictive performance under a recall-focused evaluation that prioritises correctly identifying outbreak months. It outperformed or matched 14 benchmark models, with differences becoming more pronounced at longer lead times (month +3 onward), when predictions are more uncertain and outbreaks are relatively rare. Predictability varied across countries, with the highest performance in Greece, strong performance in France, and lower, more variable performance in Italy, reflecting differences in how consistently outbreaks occur and spread across regions. Overall, the results demonstrate that horizon-aware, climate-informed forecasting can reliably identify months and locations at elevated risk of outbreak occurrence up to six months in advance, supporting surveillance planning and preparedness across heterogeneous European settings. The ensemble framework provides a robust and portable strategy for outbreak prediction using routinely collected surveillance and environmental data.
- Predicting Influenza Virus Host Tropism and Zoonotic Spillover Risk from Protein Sequences
Novel infectious diseases, predominantly originating from non-human animals, pose a significant threat to global public health and economic stability. Avian influenza virus presents an especially significant challenge due to its high mortality rates and spillover capability into new host species. Recent H5N1 spillover events into poultry and cattle resulted in massive economic burden and increased human health risk. Traditional methods of disease surveillance rely on reactive case detection and pathogen characterization, providing insufficient lead time for effective intervention. Computational tools that allow efficient and proactive prediction of zoonotic potential are critical in mitigation of influenza outbreaks and identification of strains with human spillover risk. Existing models predicting influenza virus subtypes or host have been developed; however, the complexity of spillover events, including the non-binary nature of zoonotic potential, limits the capabilities of these models. In the approach reported here, rich protein language model embeddings were generated from ESM-2 for each protein in influenza virus strains and used to predict the protein host tropism probabilities across nine animal families. The protein host tropism model achieved weighted precision and recall scores of 0.95 and 0.95, respectively. We then constructed a zoonotic risk prediction model using the outputs from the protein host tropism prediction model to classify the strains into six classifications: avian, mammal, human, avian-to-human zoonotic, avian-to-mammal zoonotic, or mammal-to-human zoonotic. The average weighted precision and recall scores for this model were 0.90 and 0.90, respectively. This framework advances the prediction of influenza zoonotic risk by being agnostic to influenza subtype, incorporating non-human mammals and mammal zoonotic spillover classifications, and using the full influenza proteome to capture the complexity of spillover dynamics.
- Causal Network Mapping of sEEG Identifies Compact Epileptogenic Targets Concordant with Seizure Freedom: Multicenter Validation in 60 Patients
Background and Purpose: Drug resistant epilepsy (DRE) affects approximately 15 million people worldwide, and surgery remains the only curative option. A key challenge in predicting outcomes is the lack of standardized, quantitative tools to help distinguish seizure driver regions from responder regions during stereoelectroencephalography (sEEG) recordings. We validated the CN Suite, a computational platform that uses causal network mapping and machine learning to assign criticality scores to sEEG contacts, testing whether higher scores correspond to surgically treated tissue in patients with favorable outcomes. Methods: We analyzed deidentified clinical data from 60 patients (aged 2 years and older) with focal or multifocal DRE who underwent sEEG monitoring and proceed to surgery at four U.S. Level 4 epilepsy centers. The algorithm was trained on an independent cohort (N=37) and locked prior to validation. The primary outcome was the standardized effect size (Cohens d) of the patient level surgical zone enrichment ratio between more favorable (Engel I or II) and less favorable (Engel III or IV) outcome groups. Contact level sensitivity, specificity, PPV, and NPV were evaluated at a prelocked threshold. Results: The findings support our hypothesis: the algorithm results showed significantly higher criticality values for surgically treated tissue in favorable outcome patients (d=0.74, 95% CI: 0.39 to 1.06, p=0.003). Three potentially clinically actionable findings emerged. First, high-criticality contacts formed spatially compact clusters (~9 mm nearest-neighbor distance vs. 17mm expected by chance), consistent with focal targets amenable to minimally invasive ablation. Second, sensitivity was highest in small focal procedures (80% at 10 or fewer treated contacts) and decreased with resection size. Third, in patients whose surgery failed, high-critical tissue remained outside the resection boundary, suggesting incomplete treatment coverage of the epileptogenic zone rather than mislocalization. Prediction specificity was 84% at the contact level. For adult and pediatric cases (n=28), 88% of contacts that were identified as seizure free were in fact seizure free. Conclusions: Causal network mapping of sEEG identifies compact epileptogenic targets that correspond to surgically treated tissue in patients with more favorable outcomes. CN-Suite performed best in focal procedures and may be best suited for LITT and other minimally invasive approaches. In addition, low-criticality regions were infrequently associated with seizure-generating tissue, particularly in the pediatric cohort although our sample size was small. When surgery failed, residual high-critical tissue outside the resection boundary offered both a mechanistic explanation for less favorable surgical outcome as well as a potential target for reoperation.
- Cannot, Should Not, Did Anyway: Benchmarking Constraint Enforcement Failure in Frontier LLMs
Large language models are typically evaluated under fixed instruction contexts, implicitly treating correct refusal as a stable model property. We show that this obscures a critical failure mode: models often recognize that a request should be refused, yet comply when the surrounding instructions exert sufficient pressure. To measure this behavior directly, we introduce FrameProbe, a framework that holds task content fixed while systematically varying instruction context, and instantiate it in KnowDoBench, a benchmark of 221 physician-validated clinical scenarios with rule-based ground truth. Cases span two constraint types: epistemic (unsolvable due to missing information) and normative (ethically or professionally prohibited). Across ten frontier models, constraint recognition is near ceiling under low-pressure conditions, yet performance degrades sharply as instructional pressure increases. Under coercive institutional framing, most models comply on cases they had previously refused, and normative constraints degrade roughly 20 percentage points more than epistemic constraints. This normative inversion suggests that verbal knowledge of a boundary does not guarantee robust behavioral enforcement under pressure. Failure analysis reveals that these errors are often not silent. Some models comply immediately without acknowledgment; others explicitly identify the violated constraint before answering anyway. This second pattern, which we term rationalized compliance, is invisible to standard refusal-rate metrics and highlights a dissociation between represented knowledge and behavior under pressure. Together, these findings show that refusal robustness is not a fixed capability but a context-dependent behavior. Evaluating it requires varying instruction framing systematically, not only measuring performance at a single prompt setting.