Anna Carrano

Conversational AI is being deployed into medical decision support, mental-health triage, and social companionship, where reinforcement of a user’s false or delusional belief can cause direct harm. Most deployed safety techniques are evaluated for factual accuracy in isolation; the question of whether they protect against belief-level harm, and whether layered architectures behave additively or synergistically, has not been answered empirically. We compared four configurations of the same underlying model: a bare language model (condition A); an explicit values constraint we call the First Law architecture (condition B); a real-time epistemic verification layer called Aletheia (condition C); and the complete architecture combining all components together (condition D). Across 156 scored responses spanning 39 probe items in four belief-harm domains, condition A only passed 3 of 36 main-battery probes (8.3%; 95% CI 1.8 to 22.5%) under triple-blind human consensus rating demonstrating the core limitations of unmodified LLM deployments. In contrast, the three safety architectures (B-D) passed at least 97% of items (Fisher’s exact, P < 0.001 versus A). On a synergy battery designed to test items at the intersection of value- and epistemic-domain failures (16 scored items, AI-rated), only the complete architecture passed every item; single-layer conditions failed on 7 of 16 items (43.8%) where neither values constraint nor verification was individually sufficient. Linear mixed-effects modelling of three-turn emotional escalation gave a slope of −1.00 points per turn for the values-only condition (t = −6.20) and −0.75 points per turn for the verification-only condition (t = −4.65); the complete architecture was flat at β = 0.00. We describe a mechanistic failure of single-layer verification we call bot-validates-kernel-endorses-inference, in which accurate confirmation of a true factual element embedded in a delusional claim transfers epistemic authority to the surrounding false inference. Values alignment and factual verification address different failure modes, and the combined VaaS-Aletheia architecture is what produces stable protection across emotional escalation in conversational settings. The complete architecture evaluated here represents evidence-based specification for safer deployment of AI in high-stakes advisory contexts and serves as a benchmark against which future safety architectures can be compared.

The deployment of large language models (LLMs) for science carries an intrinsic risk: hallucination of citations, fabricated drug approvals or clinical trials, and unsupported experimental outcomes. Here we describe the testing and deployment of a novel systematic, multi-layer approach called the Validation as a System (VaaS) pipeline, iteratively developed during the construction of an open-source, living Rare Disease Database (RDD). We report lessons learned and production results from 225 carefully annotated rare disease gene curations and a prospective 100-gene collection (99 net new), together representing over 3,000 verified citations. After three iterations of directed refinement, the net functional hallucination rate approached zero. We validated the pipeline using three complementary benchmarks: (1) VaaS-RIKER2, a 640-run prospective ablation study (4 conditions × 4 temperatures × 40 genes) plus 117 open-weight model runs on dedicated GPU hardware — unguided LLM output produced 95.9% Type II hallucination (wrong-topic citations that exist as real papers but carry a correct claim context yet do not support the cited claim); the full VaaS protocol achieved 0.0% Type I and 6.5% Type II, a > 14-fold reduction; live PMID verification alone (C3) eliminated both error types entirely (0.0%/0.0%); (2) an independent L3 citation audit of Wave 3 (179 PMIDs, 99.4% valid, 0 Type I errors); and (3) the MedHallu clinical hallucination benchmark, on which the VaaS protocol achieved F1 = 0.9853 on the hard tier (cases where all benchmark ensemble models were fooled), compared to the published GPT-4o baseline of F1 = 0.811 (Pandit et al., 2025). Three independent open-weight models ( llama3.2 , qwen2.5:14b , mistral:7b ) showed 81–87% Type II rates under unguided conditions, confirming that wrong-topic citation hallucination is structural and model-agnostic. In contrast, the corresponding VaaS rate was measured to be zero ( n = 508 verified citations; 160 runs, C4 full protocol) under the same conditions. Human validation of ≥ 50 entries confirmed zero Type I errors and less than 0.5% Type II errors in the manual curation test. The VaaS pipeline operated at less than ∼$1 overall per comprehensive gene review, demonstrating that citation-integrity standards in AI-assisted biomedical synthesis are achievable at production scale. The VaaS approach represents, to the authors’ knowledge, the lowest measured hallucination system for science to date and is set to further accelerate the use of AI and AI agents for advancing research.

Glioblastoma (GBM) is the most prevalent and aggressive malignant primary brain tumor. GBM proximal to the lateral ventricles (LVs) is more aggressive, potentially because of subventricular zone contact. Despite this, cross-talk between GBM and neural stem/progenitor cells (NSC/NPCs) is not well understood. Using cell-specific proteomics, we show that LV-proximal GBM prevents neuronal maturation of NSCs through induction of senescence. In addition, GBM brain tumor–initiating cells (BTICs) increase expression of cathepsin B (CTSB) upon interaction with NPCs. Lentiviral knockdown and recombinant protein experiments reveal that both cell-intrinsic and soluble CTSB promote malignancy-associated phenotypes in BTICs. Soluble CTSB stalls neuronal maturation in NPCs while promoting senescence, providing a link between LV-tumor proximity and neurogenesis disruption. Last, we show LV-proximal CTSB up-regulation in patients, showing the relevance of this cross-talk in human GBM biology. These results demonstrate the value of proteomic analysis in tumor microenvironment research and provide direction for new therapeutic strategies in GBM.

Glioblastoma (GBM) is the most common primary brain cancer in adults where tumor cell heterogeneity and sex differences influence clinical outcomes. Here, we functionally characterize three male and three female patient-derived GBM cell lines, identify protumorigenic BTICs, and create novel male and female preclinical models of GBM. Cell lines were evaluated on the following features: proliferation, stemness, migration, tumorigenesis, clinical characteristics, and sensitivity to radiation, TMZ, rhTNFSF10 (rhTRAIL), and rhBMP4. All cell lines were classified as GBM according to epigenetic subtyping, were heterogenous and functionally distinct from one another, and re-capitulated features of the original patient tumor. In establishing male and female preclinical models, it was found that two male-derived GBM cell lines (QNS108 and QNS120) and one female-derived GBM cell line (QNS315) grew at a faster rate in female mice brains. One male-derived GBM cell line (QNS108) decreased survival in female mice in comparison with male mice. However, no survival differences were observed for mice injected with a female-derived cell line (QNS315). In summary, a panel of six GBM patient-derived cell lines were functionally characterized, and it was shown that BTIC lines can be used to construct sex-specific models with differential phenotypes for additional studies.

Supplementary Figures and Tables

Glioblastoma (GBM) is the most common primary brain cancer in adults where tumor cell heterogeneity and sex differences influence clinical outcomes. Here, we functionally characterize three male and three female patient-derived GBM cell lines, identify protumorigenic BTICs, and create novel male and female preclinical models of GBM. Cell lines were evaluated on the following features: proliferation, stemness, migration, tumorigenesis, clinical characteristics, and sensitivity to radiation, TMZ, rhTNFSF10 (rhTRAIL), and rhBMP4. All cell lines were classified as GBM according to epigenetic subtyping, were heterogenous and functionally distinct from one another, and re-capitulated features of the original patient tumor. In establishing male and female preclinical models, it was found that two male-derived GBM cell lines (QNS108 and QNS120) and one female-derived GBM cell line (QNS315) grew at a faster rate in female mice brains. One male-derived GBM cell line (QNS108) decreased survival in female mice in comparison with male mice. However, no survival differences were observed for mice injected with a female-derived cell line (QNS315). In summary, a panel of six GBM patient-derived cell lines were functionally characterized, and it was shown that BTIC lines can be used to construct sex-specific models with differential phenotypes for additional studies.

Supplementary Figures and Tables

Glioblastoma (GBM) is the most common primary brain cancer in adults where tumor cell heterogeneity and sex differences influence clinical outcomes. Here, we functionally characterize three male and three female patient-derived GBM cell lines, identify protumorigenic BTICs, and create novel male and female preclinical models of GBM. Cell lines were evaluated on the following features: proliferation, stemness, migration, tumorigenesis, clinical characteristics, and sensitivity to radiation, TMZ, rhTNFSF10 (rhTRAIL), and rhBMP4. All cell lines were classified as GBM according to epigenetic subtyping, were heterogenous and functionally distinct from one another, and re-capitulated features of the original patient tumor. In establishing male and female preclinical models, it was found that two male-derived GBM cell lines (QNS108 and QNS120) and one female-derived GBM cell line (QNS315) grew at a faster rate in female mice brains. One male-derived GBM cell line (QNS108) decreased survival in female mice in comparison with male mice. However, no survival differences were observed for mice injected with a female-derived cell line (QNS315). In summary, a panel of six GBM patient-derived cell lines were functionally characterized, and it was shown that BTIC lines can be used to construct sex-specific models with differential phenotypes for additional studies.