Hallucination News and Research

RSS

A hallucination, in the broadest sense of the word, is a perception in the absence of a stimulus. In a stricter sense, hallucinations are defined as perceptions in a conscious and awake state in the absence of external stimuli which have qualities of real perception, in that they are vivid, substantial, and located in external objective space.

Physics-informed AI framework improves aortic aneurysm stress prediction

Ascending thoracic aortic aneurysms often progress silently, but once dissection or rupture occurs, they can cause severe mortality before hospital arrival or during admission.

4 Sep 2026

ChatGPT fills gaps like a confabulator, not a hallucinating mind

A Perspective article argues that large language model errors more closely resemble confabulations, while speech recognition errors, such as unsupported Whisper transcriptions, are functionally closer to hallucinations. The authors stress that these parallels are behavioral and provisional, not evidence of shared consciousness, and suggest that psychiatric concepts could help improve AI reliability.

11 Jun 2026

Study finds popular AI chatbots often give problematic health advice

Researchers audited five popular public-facing AI chatbots across 250 health prompts and found that 49.6% of responses were problematic, with especially weak performance on open-ended questions and in nutrition and athletic performance. The study also found poor citation quality, difficult readability, and very few refusals to answer, raising concerns about the reliability of chatbots for everyday health information.

16 Apr 2026

Can AI chatbots help brain tumor patients understand their care?

Integrating LLMs in brain tumor care could enhance patient understanding, but requires strict oversight to manage risks and ensure reliable information.

26 Mar 2026

Review explores how generative AI could support precision oncology decision-making

A narrative review examines how generative AI tools such as large language models and vision-language models could assist precision oncology by interpreting genomic data, matching patients to clinical trials, and synthesizing complex clinical information. The authors emphasize governance, human oversight, and retrieval-based safeguards to ensure these systems are deployed safely in cancer care.

11 Mar 2026

AI tool helps clinicians quickly understand complex patient medical histories

An AI-guided platform at Penn Medicine allows clinicians to quickly and easily unearth pertinent information from patients’ electronic health records that otherwise might have been difficult to find.

14 Jan 2026

Novel therapy for psychosis tackles disturbing mental imagery

A pioneering therapy targeting distressing and disturbing mental images for people with psychosis could lead to a transformative intervention for a condition estimated to cost England almost £12 billion a year.

6 Nov 2025

New AI system revolutionizes drug design and vaccine development

The rapid advancement of artificial intelligence (AI) is transforming the field of structural biology. In May 2024, DeepMind and Isomorphic Labs launched AlphaFold 3 (AF3), a revolutionary AI system capable of predicting the 3D structures and interactions of proteins, nucleic acids, small molecules, ions, and other biomolecules with remarkable precision.

10 Oct 2025

Growing use of AI scribes among New Zealand GPs raises ethical and legal questions

Many New Zealand GPs have taken up the use of AI scribes to transcribe patient notes during consultations despite ongoing challenges with their legal and ethical oversight, data security, patient consent, and the impact on the doctor-patient relationship, a study led by the University of Otago, Wellington – Ōtākou Whakaihu Waka, Pōneke has found.

14 Aug 2025

Flu virus triggers neuropsychiatric symptoms not the medication

For decades, medical professionals debated whether a common antiviral medication used to treat flu in children caused neuropsychiatric events or if the infection itself was the culprit.

4 Aug 2025

Hybrid AI approach solves problems in tractography for brain surgery

The brain is a highly complex network of nerve cells interconnected by delicate pathways - known as nerve fibers or tracts.

10 Feb 2025

GPT-4 demonstrates high accuracy in analyzing multilingual medical notes

A study evaluates GPT-4's analysis of medical notes in English, Spanish, and Italian, achieving 79% agreement with physicians on key clinical insights.

6 Jan 2025

Yale study links paranoia to visual perception deficits

Could complex beliefs like paranoia have roots in something as basic as vision? A new Yale study finds evidence that they might.

17 Dec 2024

Large language models could help automate functional genomics research

Researchers at University of California San Diego School of Medicine have demonstrated that large language models (LLMs), such as GPT-4, could help automate functional genomics research, which seeks to determine what genes do and how they interact.

2 Dec 2024

ChatGPT outperforms trainee doctors in respiratory disease assessments

The chatbot ChatGPT performed better than trainee doctors in assessing complex cases of respiratory disease in areas such as cystic fibrosis, asthma and chest infections in a study presented at the European Respiratory Society (ERS) Congress in Vienna, Austria.

9 Sep 2024