BACKGROUND:Artificial intelligence (AI) tools offer new opportunities to support human perception and provide gaze guidance in surgery. However, there remains a need to better understand surgeons' perceptions of these tools and to identify desired features across levels of expertise. This study explored surgeons' perceptions of AI-based gaze guidance to determine how such tools can best meet surgeons' needs while recognizing their limitations. METHODS:Seventy-one surgeons (38 attendings, 33 residents) watched laparoscopic cholecystectomy videos and responded to open-ended questions about their perceptions of AI for gaze guidance in surgery. Responses were analyzed using latent content analysis to identify themes and desired AI features. Each response was also categorized as positive, negative, or neutral to assess overall perception. RESULTS:Attendings (92%) and residents (82%) were positive about AI for visual gaze guidance. Neutral responses (9.8%) reflected uncertainty or limited familiarity with AI. Desired features included directing attention, identification, highlighting and visualization, feedback, decision support, educational tools, communication and collaboration, and planning and mapping. Identification involved AI support for recognizing critical structures, abnormalities, gaps in attention, and spatial orientation. Attendings favored AI for decision support and feedback, while residents emphasized its value for visualization and education. Surgeons frequently described AI as a "second pair of eyes," providing guidance, visualization, user control, and safety. Design considerations were also identified to provide guidance without being a potential distractor to the surgeon. CONCLUSION:Surgeons across expertise levels expressed positive views toward AI support for gaze guidance, emphasizing the need for features that complement human expertise, align with individual training needs, and account for contextual factors such as minimizing distractions and preventing tunnel vision. Identification, direct attention, and decision support were the most desired features. The findings suggest that tailored AI-based tools can play an important role in surgical training and active procedures by supporting harm reduction.
The deployment of decision-making AI agents presents a critical challenge in maintaining alignment with human values or guidelines while operating in complex, dynamic environments. Agents trained solely to achieve their objectives may adopt harmful behavior, exposing a key trade-off between maximizing the reward function and maintaining alignment. For pre-trained agents, ensuring alignment is particularly challenging, as retraining can be a costly and slow process. This is further complicated by the diverse and potentially conflicting attributes representing the ethical values for alignment. To address these challenges, we propose a test-time alignment technique based on model-guided policy shaping. Our method allows precise control over individual behavioral attributes, generalizes across diverse reinforcement learning (RL) environments, and facilitates a principled trade-off between ethical alignment and reward maximization without requiring agent retraining. We evaluate our approach using the MACHIAVELLI benchmark, which comprises 134 text-based game environments and thousands of annotated scenarios involving ethical decisions. The RL agents are first trained to maximize the reward in their respective games. At test time, we apply policy shaping via scenario-action attribute classifiers to ensure decision alignment with ethical attributes. We compare our approach against prior training-time methods and general-purpose agents, as well as study several types of ethical violations and power-seeking behavior. Our results demonstrate that test-time policy shaping provides an effective and scalable solution for mitigating unethical behavior across diverse environments and alignment attributes.
Quantitative kinematic tracking of in vivo skeletal structures efficiently and with high temporal resolution is important to a wide range of questions in biomechanical research. As such, software tools are needed for efficient, semi-automated registration of bones imaged with either 3DCT or 4DCT. This study presents Hierarchical 3D Registration (3DH), an open-source approach to skeletal tracking of three-dimensional image volume sets. The 3DH approach and software are presented using sequential 3DCT datasets of participants performing various thumb tasks and dynamic 4DCT datasets of simulated flexion–extension in cadaveric wrists collected in previous studies. The agreement in computed arthrokinematics with previously calculated values using independent approaches for 3DCT and 4DCT data was assessed with Bland–Altman analyses. Using 3DH, all target bones were successfully tracked for both the 3DCT and 4DCT sets. For 3DCT data of the radius, first metacarpal and trapezium, the mean bias for the helical angle was −0.06 degrees with 95
Pathology remains central to clinical diagnosis, yet adoption of digital pathology is constrained by financial, operational, and workflow burdens of fully digital infrastructure. We introduce HistoCAM, a platform for ambient, real-time datafication and digitization of glass-slide microscopy that preserves microscope workflows. A 31-megapixel, high space-bandwidth-time-product camera and custom software application stream and composite the pathologist’s eyepiece view, passively generating multi-resolution images from 2X to 40X while recording magnification use, search paths, and dwell times. These outputs provide immediate workflow uplift through digital annotation, measurement, quality assurance, and real-time integration of configurable AI tools. Simultaneously, HistoCAM links image content with expert interaction data and supports rapid generation of annotated, pre-embedded training data during routine slide review. By converting routine microscopy into an AI-ready data stream without requiring additional acquisition steps, HistoCAM provides a practical bridge to computational pathology while creating process-aware datasets that capture how pathologists examine and interpret tissue. HistoCAM ambiently captures the pathologist’s full, optically sampled microscope view, enabling real-time digital and AI tools while turning routine slide review into process-aware, AI-ready data.
The phase stability, crystal structure, and melting of Nb have been examined under high pressure shock compression to 365 GPa using ultrafast x-ray diffraction measurements on two x-ray free electron laser facilities. On compression, Nb remains stable in the bcc phase up to 220 GPa, with coexistence of bcc-Nb and liquid from 249 to 298 GPa, and complete melting at 301 GPa, with melt identified by diffuse liquid diffraction. Melting initiates at higher pressure than expected based on theoretical predictions, and the data consistently excludes the presence of other suggested phases of Nb at these pressures, including and hcp, thereby resolving the long-standing structural discrepancy. Singh-type strength analysis provides experimental evidence for a change in sign of the elastic anisotropy parameter at 121 GPa, consistent with the previously reported heat-induced hardening to heat-induced softening transition in Nb.
