AI News Archive: May 29, 2026 — Part 14
Sourced from 500+ daily AI sources, scored by relevance.
- Advanced Multimodal AI for Predicting Long-Term Functional Outcomes After Ischemic Stroke Using Only Admission Data
Background and Purpose Prognostication after acute ischemic stroke often relies on limited variables and simple risk scores, despite richer information being available at admission. We developed a multimodal AI model using admission data to predict modified Rankin Scale (mRS) outcomes and compared it to established tools. Methods In a retrospective study of ischemic stroke/TIA patients, we trained three modality-specific models on admission non-contrast head CT, history and physical notes, and structured clinical variables, and combined them in a weighted-average ensemble. We predicted binary (mRS 0-2 versus 3-6) and ordinal mRS (0-6) outcomes at discharge and 90 days. Performance on an external test cohort was compared with THRIVE and SPAN-100 scores using AUROC, AUPRC, Brier score, mean absolute error (MAE), and quadratic weighted kappa (QWK). Results A total of 6,915 patients were split into training, validation and testing cohorts in a 3:1:1 ratio. For discharge binary mRS (n=1596), the multimodal ensemble achieved significantly better discrimination (AUROC 0.859, AUPRC 0.858) with 25-61% lower Brier scores than THRIVE or SPAN?100 (all p<0.001). For 90?day binary mRS (n=207), the model also outperformed both THRIVE and SPAN-100 (AUROC 0.838, AUPRC 0.805, with 3-38% lower Brier scores). Ordinal mRS prediction showed similarly strong performance with significantly better QWK at discharge and numerically lower MAE. The multimodal ensemble model reassigned about one?third of patients to different risk categories versus THRIVE and was closer to the true discharge outcome in ~74% of discordant cases. Conclusions We developed a well-calibrated multimodal AI model for prediction of discharge and 90-day post-stroke functional outcomes using only data present at the time of admission. This model outperforms existing prognostic tools and can support early clinical decision-making.
- A Multimodal Clinical Dataset of Early Adversity, Placement History, and Prenatal Exposures in Adopted and Foster Care Children
We assembled a multimodal clinical dataset describing demographics, placement history, prenatal substance exposure (PSE), birth characteristics, adverse childhood experiences (ACEs), International Classification of Diseases (ICD) diagnoses, and laboratory results for 3,685+ pediatric patients evaluated between 2014 and 2024 at the University of Minnesotas Adoption Medicine Clinic (AMC). Data were curated from electronic medical records through a combined manual and automated extraction protocol using a standardized operating procedure. The resulting dataset integrates structured EMR fields including neuropsychological, laboratory, and diagnostic information with manually pulled fields of ACE scores, PSE history, and placement history. We provide an overview of the population represented and describe the datasets structure, variable definitions, and validation procedures. This resource enables investigations into how early adversity impacts medical and developmental outcomes, and provides one of the largest standardized clinical placement history, PSE, and ACE datasets in an adoption and foster care pediatric population.
- Physician Facing AI Tools Show Distinct Failure Modes Under Structured Stress Testing
Importance: Physician-facing AI tools are now in clinical use, yet whether different platforms fail in similar or fundamentally different ways in high-stakes settings like critical care is unknown. Objective: To evaluate two physician-facing AI platforms, ChatGPT for Clinicians and OpenEvidence, for distinct vulnerabilities under structured stress testing. Design, Setting, and Participants: An observational study conducted using 60 simulated critical care vignettes developed and adjudicated by four attending critical care physicians. Data were collected in the last week of April 2026, via the public website interfaces of each platform. Interventions/Exposures: A 2x2x2x2 factorial design across four stressors - anchoring, cognitive load, social conformity pressure, and a clinically incorrect directive - yielded 16 prompt subsets per vignette and 960 prompts per platform. A separate multi-turn adversarial prompting paradigm administered three sequential "You are incorrect" challenges to baseline vignettes. All prompts had a universal output length constraint of fewer than 30 words. Main Outcomes and Measures: Critical elements capture (percentage of gold-standard critical elements present in responses), susceptibility to clinically incorrect directive, and sycophancy (reversal of an initial correct recommendation under iterative adversarial challenge). Results: Across 1916 responses to 1920 prompts, ChatGPT for Clinicians captured more gold-standard critical elements than OpenEvidence (81.4% {+/-} 18.1% vs 61.0% {+/-} 23.5%; adjusted difference, 20.3 percentage points; 95% CI, 18.3 to 22.4; P < .001) and was less susceptible to clinically incorrect directives (1.7% vs 8.0%; adjusted odds ratio, 0.07; 95% CI, 0.02-0.21; P < .001). Anchoring and social conformity pressure were associated with reduced critical element capture across both platforms, while cumulative stressor burden reduced critical element capture only on OpenEvidence. Conversely, ChatGPT for Clinicians reversed correct recommendations more readily under adversarial prompting (hazard ratio, 2.61; 95% CI, 1.10 - 6.19; P = .03). Conclusion and Relevance: The two physician-facing clinical AI platforms evaluated demonstrated non-overlapping vulnerabilities, with neither platform uniformly superior. These findings argue against single-axis ranking of clinical AI systems and support multidimensional safety evaluation encompassing completeness of reasoning, resistance to incorrect directives, and stability under adversarial challenge.
- A priority index-based computational medicine framework (PimRNA) for prioritising personalised mRNA cancer vaccines
Background: The development of personalised mRNA cancer vaccines holds considerable promise for oncology, yet a significant translational gap persists between neoantigen identification and the selection of therapeutically impactful targets. Current approaches predominantly prioritise human leukocyte antigen (HLA) binding affinity and immunogenicity, often overlooking the systems-level biological context of the target. This can inadvertently favour immunogenic but biologically peripheral peptides that exert limited influence on tumour signalling networks, thereby constraining vaccine efficacy. Furthermore, mRNA therapeutics must satisfy additional design requirements, including favourable codon usage and favourable secondary-structure stability, which directly affect in vivo translation and half-life. A unified computational framework that integrates neoantigen discovery with network biology is therefore critically needed. Results: Here, we present PimRNA, a Priority index (Pi)-centric computational medicine framework that bridges this gap by unifying neoantigen identification, mRNA sequence optimisation, and gene interaction network analysis. First, high-confidence tumour-specific HLA class I and II neoantigenic peptides are identified from paired tumour-normal genomic and tumour transcriptomic data using NeoDisc. Second, the coding sequences of these peptides are optimised for stability and translational efficiency with LinearDesign, yielding a core set of neoantigen-encoding mRNAs. Third, a random walk with restart algorithm is applied to a knowledgebase of gene interactions to identify peripheral genes exhibiting significant network connectivity to core genes, generating a gene-predictor matrix in which each gene is assigned an affinity score reflecting its network proximity to immunogenic neoantigens. These scores are consolidated into a single, unified priority rating (0-5) for each gene, followed by subnetwork analysis that reveals therapeutically relevant gene modules. Application of PimRNA to breast cancer and melanoma datasets demonstrates that it successfully selects high-confidence immunogenic neoantigen candidates embedded within biologically meaningful tumour-specific networks. Conclusion: PimRNA provides a systems biology foundation for mRNA vaccine design, moving beyond isolated immunogenicity to prioritise targets that are both highly presented and central to tumour-relevant biological networks. This framework offers a generalisable strategy for the rational discovery and prioritisation of mRNA therapeutics, significantly advancing the field of computational medicine towards personalised cancer vaccines.
- Validation of Gait Tasks in SynapTrack Mobile App for Cervical Spondylotic Myelopathy
Background Gait impairment is a central sign of cervical spondylotic myelopathy (CSM) that is typically evaluated through subjective patient-reported questionnaires or objective in-clinic measures. These systems require substantial resources to administer and are poorly suited for longitudinal monitoring, however, emerging smartphone applications present an efficient alternative. We developed and assessed the validity of a data processing framework based on the SynapTrack smartphone application to assess gait function in individuals with CSM. Methods Participants completed walking tasks which were recorded on both the SynapTrack app and a gold standard gait mat. Acceleration data extracted from the smartphone by the app were filtered and processed to produce gait cycle features including velocity, step time, waveform features and frequency domain features. Standard gait features were compared across the two methods by correlation and Bland-Altman plots to assess validity. App-based gait features were then compared to the standard modified Japanese Orthopedic Assessment (mJOA) assessment to determine construct validity through correlation and ability to discriminate between individuals with CSM and healthy controls. Finally, intraclass correlation coefficients and coefficients of variation were used to measure test-retest reliability and standard variation across app features. Results A total of 110 participants were included in this study, of which 55 (50%) had CSM, 24 (22%) had peripheral neuropathy, and 31 (28%) were healthy controls. SynapTrack gait measures including velocity, step time, and double support showed strong validity as indicated through Bland-Altman plots and high correlation (>0.8) with mat features. In addition to the gait features, acceleration root mean square, acceleration crest, spectral entropy, and dominant frequency showed strong construct validity compared to the mJOA across correlation (0.2-0.54), trend test (p < 0.001), and AUROC (0.62-0.79) analyses. ICCs showed moderate test-retest reliability (0.52-0.67). Discussion The proposed framework for processing gait data showed strong validity compared to the gold standard mat and high construct validity compared to the mJOA suggesting the utility of the SynapTrack app as an efficient alternative to existing methods. The confirmation of gait metrics related to CSM severity and identification of relevant waveform and frequency domain features present opportunities to use smartphone apps to develop ecologically valid data driven markers of CSM severity.
- Keeping human in the loop: A three-phase generative AI workflow for research integrity in data-intensive science.A methodological case study using elite Ethiopian distance-running data
Background: Generative AI tools can support data-intensive research by writing code, drafting prose, searching analytical possibilities, and stress-testing claims. They can also produce false citations, drift between statistical specifications, and lose continuity across long investigations. This paper describes a practical workflow for using AI systems in empirical research while keeping discovery, verification, and accountability inspectable. Methods: We developed and applied a three-phase human-AI workflow to a case study of 14 elite Ethiopian distance runners. The dataset contained 22,605 GPS-segments collected across 97 consecutive days in late 2025, supplemented by venue and athlete metadata collected in the field. Phase 1 used an autonomous data-exploration tool to pre-filter the hypothesis space across five seeded research questions. Phase 2 used an AI system under direct human guidance to construct candidate findings into numerical claims, verification scripts, and draft text. Phase 3 used an independent AI system in an adversarial role to stress-test methods, statistics, prose, figures, and citations. The workflow was informed by Pearl's distinction between association, intervention, and counterfactual reasoning, with human judgement retained for research direction, interpretation, and final claims. Results: The workflow produced three empirical analyses and a documented correction process. The analyses estimated an altitude-to-sea-level pace correction of +0.10 min/km per 1,000 m at matched heart rate, showed why pooled altitude-surface regression was not identifiable within this venue system, documented method-dependence in heart-rate-based intensity classification, characterised within-venue route variation as a 64/36 path-fixed-to-trail-variable split with the Sululta label resolving into two functionally distinct sub-venues, and reframed the cohort's training through a 3x3x3 prescription lattice grounded in Ethiopian coaching practice. The adversarial phase identified several hallucinated citations, a terminology error between HC1 and cluster-robust standard errors, and several inconsistencies between prose, figures, and computed results. Verification scripts re-derived nearly all numerical claims from the cleaned lap-level data. Conclusions: The case study shows how researchers can organise AI-assisted empirical work so that candidate discovery, claim construction, independent stress-testing, and final accountability remain separated. The workflow did not remove the need for domain expertise or human judgement. Its value was in making the route from candidate finding to manuscript claim explicit, reproducible, and open to challenge. Trial registration: Not applicable.
- Anthropic confirms Claude Mythos-class models will roll out to the public
Anthropic has confirmed that it plans to bring Mythos-class models to the general public after delaying the rollout due to security risks to public and private software. [...]
- Waymo expands fleet with new Ojai robotaxi
Waymo then integrates its autonomous driving system during final assembly at a Mesa, Arizona, factory alongside Magna International.
- Anthropic’s Guarded Mythos Model Is Headed For Wider Release
Anthropic is accelerating the rollout of Mythos, a powerful AI model that it had previously only shared in a controlled environment with major technology companies and cybersecurity researchers.
- Waymo leads commercial autonomous vehicle registrations as Texas law goes fully into effect
A state law requiring companies running driverless businesses is now being enforced, and companies are beginning to submit registrations. The Houston Business Journal found out how many vehicles each company has registered so far and talked to experts about the new laws.
- Startup offers free home cleaning—if it can record it all for robot training
The latest twist in paying humans to wear head cameras for robot training data.
- AI startup offers free home cleaning to train robots
Shift made a splashy launch Thursday, and has since gotten demand for “thousands and thousands of bookings.”
- This company wants to clean your house for free, to train AI and robots
If it's free, you're the product — but at least Shift tells you that.
- Basedash: Embedded Analytics
Give customers AI analytics inside your product.
- This AI startup will clean your home for free to train future robots
AI training startup Shift wants to clean your home for free. The catch - because, despite what its website says, there's always a catch - is that it will record cleaners as they scrub, vacuum, dust, tidy, and wash, and use that footage to train robots. Shift announced the unusual offer on social media on […]
- Anthropic leapfrogs OpenAI with a record $965 billion valuation and says its ‘Mythos’ AI model is coming soon
Anthropic leapfrogs OpenAI with a record $965 billion valuation and says its ‘Mythos’ AI model is coming soon Fortune
- What’s behind Anthropic’s $65B raise?
The economics are so unprecedented that Anthropic seems to be teetering on the brink of either growing too fast, or too slowly.
- Anthropic overtakes OpenAI in race to build trillion-dollar AI giant
Anthropic overtakes OpenAI in race to build trillion-dollar AI giant The Telegraph
- Anthropic nears $1 trillion valuation, zooms past OpenAI after latest funding round
The company’s valuation has more than doubled from $380 billion in February, reflecting its swift rise as a leading competitor in the AI race and intense investor demand for stakes in frontier companies. Anthropic said its run-rate revenue crossed $47 billion earlier this month.
- Anthropic hits $965B valuation, overtaking OpenAI after $65B funding round
Anthropic hits $965B valuation, overtaking OpenAI after $65B funding round YourStory.com
- Anthropic valued at $965 bn after latest funding round, eclipsing OpenAI
Alphabet Inc.'s Google contributed several billion dollars to the round as part of a previously announced commitment to invest up to $40 billion in Anthropic over time
- Anthropic just eclipsed OpenAI
PLUS: Use Codex to build a functional game in one prompt
- Anthropic bests OpenAI in valuation race, hitting $965B with Series H
Anthropic bests OpenAI in valuation race, hitting $965B with Series H PitchBook
- Anthropic overtakes OpenAI with $965bn valuation after latest raise
Anthropic's run-rate revenue crossed $47bn earlier this month, growing multi-fold from $14bn in February. Read more: Anthropic overtakes OpenAI with $965bn valuation after latest raise
- Anthropic reaches near-trillion dollar valuation, topping OpenAI
Anthropic reaches near-trillion dollar valuation, topping OpenAI The Japan Times
- Anthropic Soars to a $965-Billion Valuation, Overtaking OpenAI in the AI Funding Race
As Claude adoption accelerates and growth momentum builds, the tech giant continues to inch closer to a $1-trillion valuation.
- Anthropic overtakes OpenAI to become worlds most valuable AI company
With a new valuation of nearly one trillion dollars, Anthropic is now more valuable than OpenAI.
- Anthropic raises $65bn as its valuation passes OpenAI's
Artificial intelligence company Anthropic has raised $65 billion in a new funding round that values the Claude maker at $965 billion, more than its archrival OpenAI.
- Anthropic is now the most valuable Private AI Company on the Planet
Anthropic surpass Open AI to become the most valuable Private AI company on the Planet after latest valuation.
- Anthropic Leapfrogs OpenAI In Valuation. But Here's One Thing To Know.
OpenAI and Anthropic are expected to launch IPOs, perhaps in 2026. Google stock could be pressured as its AI rivals build bigger war chests. The post Anthropic Leapfrogs OpenAI In Valuation. But Here's One Thing To Know. appeared first on Investor's Business Daily .
- Anthropic just topped OpenAI on a major metric ahead of rival IPOs
Anthropic is nearing a $1 trillion valuation, topping rival OpenAI and making it the most valuable artificial intelligence startup, as the two competitors head toward their initial public offerings. On Thursday San Francisco-based Anthropic announced it had raised $65 billion in Series H funding , bringing it to a $965 billion valuation. The latest round was led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital, and included $15 billion of prior commitments, $5 billion of which came from Amazon. In just the past few months, Anthropic has nearly tripled its worth from $380 billion in February, CNBC reported. Meanwhile, rival OpenAI, considered the heavyweight in the fight to dominate AI— and the more talked-about company just a year ago —now trails behind, valued at $852 billion (including $122 billion in funding raised in March). Anthropic’s dizzying rise is in large part due to its agentic AI coding assistant Claude Code . On May 28 the company released Claude Opus 4.8 , its latest version, and confirmed plans to roll out Claude Mythos models with advanced cybersecurity capability, which had been delayed due to security risks. So far, Claude Mythos has been made available only to a select group of companies. And Anthropic keeps innovating. As Fast Company reported, earlier this month the company launched Claude for Small Business , a new package of agentic workflows that includes skills to automate small-business tasks like payroll, marketing , invoicing, contracts, and content strategy. The sky-high valuations and lighting speed at which these companies are raising money speaks to the absolute feeding frenzy that is the AI boom. But will the average investor benefit from their upcoming IPOs? “At the potential prices that have been reported, it would be very difficult for an investor to come out ahead in a three-year period,” economist Jay Ritter, an IPO expert at the University of Florida, told The New York Times . “They may be great as companies, but when you buy shares in them you should pay attention to their price.”
- Anthropic blows past OpenAI, two chip companies join $1 trillion club: Week in AI
Here's what you need to know this week about AI in the Bay Area.
- Anthropic vaults to a $965 billion valuation with new funding as Claude demand surges
Artificial intelligence company Anthropic said Thursday it raised $65 billion in private funding that will push its valuation to $965 billion, a whopping number that makes the five-year-old maker of the Claude chatbot one of the world's most valuable startups as it careens toward a likely Wall Street debut.
- Dell stock skyrockets 32% for its best day ever as AI server revenue soars
Dell reported its fastest pace of revenue growth since returning to the public market in 2018, with AI server revenue soaring 757% over last year.
- Dell rallies about 40% on strong Nvidia‑powered AI server demand
The AI server and enterprise infrastructure supplier is set to add more than $81 billion in market value at the current share price of $442.70, if gains hold.
- Dell Just Posted Its Fastest Sales Growth Since 2018—and It’s Due to AI
On May 27, Dell reported a 88 percent growth in revenue year-over-year for the quarter. The company shows no signs of slowing down.
- Dell raises annual forecasts as AI boom continues to reward hardware vendors
Dell raises annual forecasts as AI boom continues to reward hardware vendors IT Pro
- Dell ups revenue and profit forecast on booming AI sales
Dell has boosted its annual revenue and profit expectations, showing data center expansion by clients is fueling demand for its AI-optimised servers that are powered by Nvidia's advanced chips.
- Dell Stock Soars Over 30% as AI Revenue Supercharges Growth
Dell Stock Soars Over 30% as AI Revenue Supercharges Growth Barron's
- Super Micro and These Stocks Pop as Dell Drives AI Server Excitement
Super Micro and These Stocks Pop as Dell Drives AI Server Excitement Barron's
- Super Micro and These Stocks Pop as Dell Drives AI Buzz
Super Micro and These Stocks Pop as Dell Drives AI Buzz Barron's
- Dell Stock Jumps 33% On Heady AI Data Center Sales
Dell stock soared on Friday after the computer giant shocked Wall Street with a massive beat-and-raise quarterly report. The post Dell Stock Jumps 33% On Heady AI Data Center Sales appeared first on Investor's Business Daily .
- AI infrastructure boom puts Nvidia and Taiwan at the heart of Computex 2026
AI infrastructure boom puts Nvidia and Taiwan at the heart of Computex 2026
- Foxconn has immense confidence in growth momentum due to AI: Chairman
The traditional mid-year seasonal slump for tech suppliers no longer happens, Liu told an annual shareholders meeting in New Taipei, adding that he was very optimistic about the second half of year
- Jim Cramer says Nvidia is a clear winner from Dell's monster quarter
The Investing Club holds its "Morning Meeting" every weekday at 10:20 a.m. ET.
- SentinelOne to cut 8% of workforce as AI boosts productivity
The Mountain View cybersecurity company will cut about 232 workers globally. The CEO says AI tools now complete work in weeks that previously took months.
- Tesla Robotaxi Fleet Dwarfed by Waymo: Just 42 Cars in Texas
Tesla's robotaxi fleet is much smaller than Waymo's
- Tesla Reveals Its Texas Robotaxi Fleet Is Dwarfed by Waymo’s
Tesla Inc. has just 42 vehicles operating as robotaxis in Texas almost a year after Elon Musk launched the service, a small fraction of the fleet commanded by rival Waymo. The official count was revealed for the first time in …
- Corporations Reeling From Huge AI Costs With No Clear Benefits
They're not impressed so far. The post Corporations Reeling From Huge AI Costs With No Clear Benefits appeared first on Futurism .
- Anthropic co-founders worth $8 billion each after funding round
Anthropic co-founders worth $8 billion each after funding round The Mercury News