Browsing by Subject "Generative AI"

Now showing 1 - 5 of 5
  • Thumbnail Image
    Item
    Generative AI in STEM Teaching: Opportunities and Tradeoffs
    (2025) Price, Jeremy; Grover, Shuchi
  • Thumbnail Image
    Item
    Public knowledge of food poisoning, risk perception and food safety practices in Saudi Arabia: A cross-sectional survey following foodborne botulism outbreak
    (Wolters Kluwer, 2025) Alhuzaimi, Abdullah; Aljamaan, Fadi; Al-Ajwad, Fatema H.; Alabdulkareem, Wejdan; Alshahrani, Fatimah S.; Altamimi, Ibraheem; Al-Eyadhy, Lama; Bukhari, Abdullah; BinOmair, Alanoud; Al-Subaie, Sarah; Shayah, Yamen; Alhaboob, Ali; Alanteet, Alaa A.; Alanteet, Abeer; Alharbi, Mohammad; Jamal, Amr; Barry, Mazin; Assiri, Rasha Assad; Alhasan, Khalid; Al-Tawfiq, Jaffar A.; Temsah, Mohamad-Hani; Medicine, School of Medicine
    To investigate food poisoning knowledge, risk perception and safe food handling practices among Saudi Arabian public following foodborne botulism outbreak. A cross-sectional survey targeting the Saudi Arabian public between May 6 to 20, 2024, following the first foodborne botulism outbreak. Infectious disease and public health experts developed survey questions according to Saudi Public Health Authority and Ministry of Health (MOH) guidelines, and distributed surveys through social media. Of 3779 participants, 73.1% were female and 50.1% were aged 18 to 24 years. Almost one-third (30.2%) reported a previous food poisoning experience, with an incidence of 71.7 cases per 1000 person years. The most common perceived source of FP was restaurants foods (80.3%). The overall knowledge score of the participants regarding food poisoning was 3.42 ± 1.57 out of 7. The mean food safety practice score was 3.70 ± 1.42 out of 9. Multivariable regression analysis showed individuals aged 35 years or older (β = 0.205, P < .001), those who were married (β = 0.204, P = .003), participants with previous (FP) experience (β = 0.089, P = .009), and those who relied on information from the Ministry of Health or medical publications regarding FP (P < .001) exhibited significantly higher practice scores than other groups. The least adherence to safe practices were noted among the following: routine use of thermometer during cooking (2.7%), avoidance of washing raw chicken (13.7%) and washing hands after using cellphone during cooking (26.1%). The FP knowledge score did not correlate significantly with practice score (P = .065). This study highlights the significant knowledge gaps and inadequate food safety practices among the public in Saudi Arabia. Although certain groups, including adults (>35 years), married individuals, and those with previous food poisoning experience, showed greater adherence to safe food handling practices, adherence to specific preventive measures remained generally low. These findings highlight the need for targeted educational initiatives and interventions to improve food safety awareness and practices across diverse demographic groups in Saudi Arabia. The integration of generative AI tools, such as ChatGPT, as a public resource for food poisoning information, presents a new opportunity, but it requires further research and development to ensure accuracy and reliability.
  • Thumbnail Image
    Item
    RAPID: DRL-AI: Investigating A Community-Inclusive AI Chatbot to Support Teachers in Developing Culturally Focused and Universally Designed STEM Activities
    (2024-09-14) Price, Jeremy; Chakraborty, Sunandan
    Research to uncover and build out the initial feature set for a generative AI chatbot to support teachers in developing more culturally responsive and sustaining STEM lesson plans and activities.
  • Thumbnail Image
    Item
  • Thumbnail Image
    Item
    Zero-shot Learning with Minimum Instruction to Extract Social Determinants and Family History from Clinical Notes using GPT Model
    (IEEE, 2023) Bhate, Neel Jitesh; Mittal, Ansh; He, Zhe; Luo, Xiao; Computer Science, Luddy School of Informatics, Computing, and Engineering
    Demographics, social determinants of health, and family history documented in the unstructured text within the electronic health records are increasingly being studied to understand how this information can be utilized with the structured data to improve healthcare outcomes. After the GPT models were released, many studies have applied GPT models to extract this information from the narrative clinical notes. Different from the existing work, our research focuses on investigating the zero-shot learning on extracting this information together by providing minimum information to the GPT model. We utilize de-identified real-world clinical notes annotated for demographics, various social determinants, and family history information. Given that the GPT model might provide text different from the text in the original data, we explore two sets of evaluation metrics, including the traditional NER evaluation metrics and semantic similarity evaluation metrics, to completely understand the performance. Our results show that the GPT-3.5 method achieved an average of 0.975 F1 on demographics extraction, 0.615 F1 on social determinants extraction, and 0.722 F1 on family history extraction. We believe these results can be further improved through model fine-tuning or few-shots learning. Through the case studies, we also identified the limitations of the GPT models, which need to be addressed in future research.