GenAI - DSCI 321 - Health Analytics Library Research Guide

Generative Pre-trained Transformer (GPT)

The real skill you will need to use GenAI is knowing how to Prompt (explain) to the system what you want it to do.

This will take:

Time
- To learn to write prompts that provide the information you need
- You may have to run multiple iterations of your query.
- Then, you will have to develop a research plan to identify and properly cite the information provided by the GenAI response.
- You will need an above-average knowledge of medical and healthcare words and terminology to get proper value from using GenAI,
Money
- The free versions of most GenAI applications are 'loss leaders' and are designed to interest you. For example, ChatGPT3 is free but limits you to the number of Prompts you can submit daily.
- ChatGPT4 is a fee based service
- The cost of using a GenAI applications total rest with the user
Using PICO(T) will help with writing better GenAI Prompts

Prompt Engineering as an Important Emerging Skill for Medical Professionals: Tutorial
Prompt engineering is a relatively new field of research that refers to the practice of designing, refining, and implementing prompts or instructions that guide the output of large language models (LLMs) to help in various tasks. With the emergence of LLMs, the most popular one being ChatGPT that has attracted the attention of over a 100 million users in only 2 months, artificial intelligence (AI), especially generative AI, has become accessible for the masses. This is an unprecedented paradigm shift not only because of the use of AI becoming more widespread but also due to the possible implications of LLMs in health care. As more patients and medical professionals use AI-based tools, LLMs being the most popular representatives of that group, it seems inevitable to address the challenge to improve this skill. This paper summarizes the current state of research about prompt engineering and, at the same time, aims at providing practical recommendations for the wide range of health care professionals to improve their interactions with LLMs.
More about this database...less...
Metaaata for creating an APA 7 citation for this peer-reviewed journal article.

Meskó, B. (2023). Prompt Engineering as an Important Emerging Skill for Medical Professionals: Tutorial. Journal of Medical Internet Research, 25(6). https://doi.org/10.2196/50638

GenAI hallucination is a phenomenon in which the large language model (LLM)—often a generative model—is inaccurate or made up.

The Healthcare sector in B.C. is a highly regulated area. The College of Physicians and Surgeons of British Columbia (CPSBC) released its Interim Guidance—Ethical Principles for Artificial Intelligence in Medicine, and the British Columbia College of Nurses and Midwives (BCCNM) has also adopted protocol for AI in their clinical practices.

The American Nurses Association (ANA) released a position statement on "The Ethical Use of Artificial Intelligence in Nursing Practice" in 2022. This report identified some critical considerations when working with AI.

Alkaissi, H., & McFarlane, S. I. (2023). Artificial Hallucinations in ChatGPT: Implications in Scientific Writing. Cureus, 15(2), e35179. https://doi.org/10.7759/cureus.35179
While still in its infancy, ChatGPT (Generative Pretrained Transformer), introduced in November 2022, is bound to hugely impact many industries, including healthcare, medical education, biomedical research, and scientific writing. Implications of ChatGPT, that new chatbot introduced by OpenAI on academic writing, is largely unknown. In response to the Journal of Medical Science (Cureus) Turing Test - call for case reports written with the assistance of ChatGPT, we present two cases one of homocystinuria-associated osteoporosis, and the other is on late-onset Pompe disease (LOPD), a rare metabolic disorder. We tested ChatGPT to write about the pathogenesis of these conditions. We documented the positive, negative, and rather troubling aspects of our newly introduced chatbot's performance.
Careless Whisper: Speech-to-Text Hallucination Harms
Speech-to-text services aim to transcribe input audio as accurately as possible. They increasingly play a role in everyday life, for example in personal voice assistants or in customer-company interactions. We evaluate Open AI’s Whisper, a state-of-the-art automated speech recognition service outperforming industry competitors, as of 2023. While many of Whisper’s transcriptions were highly accurate, we find that roughly 1% of audio transcriptions contained entire hallucinated phrases or sentences which did not exist in any form in the underlying audio. We thematically analyze the Whisper-hallucinated content, finding that 38% of hallucinations include explicit harms such as perpetuating violence, making up inaccurate associations, or implying false authority. We then study why hallucinations occur by observing the disparities in hallucination rates between speakers with aphasia (who have a lowered ability to express themselves using speech and voice) and a control group. We find that hallucinations disproportionately occur for individuals who speak with longer shares of non-vocal durations—a common symptom of aphasia. We call on industry practitioners to ameliorate these language-model-based hallucinations in Whisper and to raise awareness of potential biases amplified by hallucinations in downstream applications of speech-to-text models.
Mendel and UMass Amherst Unveil Groundbreaking Research on AI-driven Hallucination Detection in Healthcare
SAN JOSE, Calif. and AMHERST, Mass., Aug. 6, 2024 /PRNewswire/ -- Mendel, a leader in healthcare AI, and the University of Massachusetts Amherst (UMass Amherst) have jointly published pioneering research addressing the critical issue of faithfulness hallucinations in AI-generated medical summaries. This collaborative effort marks a significant advancement in ensuring the safety and reliability of AI applications in healthcare settings.
More about this database...less...
Metadata to create APA 7 Citation:

Mendel and UMass Amherst Unveil Groundbreaking Research on AI-driven Hallucination Detection in Healthcare. (2024, August 6). PR Newswire, NA. https://link-gale-com.okanagan.idm.oclc.org/apps/doc/A803853923/ITBC?u=kelo91364&sid=ebsco&xid=445664cb
Faithfulness Hallucination Detection in Healthcare AI
General-purpose GenAI tools such as ChatGPT can generate texts about medical contents, pass medical exams, and perform remarkably well on multiple healthcare and medical tasks. They can summarize, abstract patient information, and answer questions about medical records, demonstrating their potential to assist medical professionals on mundane and time-consuming problems. However, there their creation of hallucinations remains problematic. While in some cases, such hallucinations might be inconsequential, in other cases, they pose significant risks in healthcare applications where soundness and trustworthiness are crucial. Inaccuracies in high-stakes healthcare settings can lead to severe consequences, including misdiagnoses and inappropriate treatments. A critical challenge that must be solved to use AI in healthcare more reliably is eliminating such hallucinations. This conference paper addresses this issue,
More about this database...less...
Metadata to help with the creation of APA 7 citation:

Rumale, P. Tiwari, S., Naik, T. G., Gupta S., Thai, D. N., Zhao, W., Adrdulov, V., Tarabishy, K., McCallum, A., & Salloum. W.. (2024). Faithfulness Hallucination Detection in Healthcare AI. KDD-AIDSH 2024, Aug 26, 2024, Barcelona, Spain. https://openreview.net/pdf?id=6eMIzKFOpJ

GenAI uses LLMs that are designed to predict the next best possible word based on their training data set; there is always the possibility of bias in the results produced by GenAI tools due to the inherent bias in the data used to build it. You should constantly closely scrutinize any results from GenAI for potential bias in the text it provides you.

REMEMBER GENAI HAS BEEN PRIMARILY TRAINED ON DATA FROM THE WEB

Bias to be aware of:

Under-representation of specific people or groups
Discrimination
Misinformation and Disinformation
Denialism
Plagiarization
Research Fraud

Bias in AI-based models for medical applications: challenges and mitigation strategies
Artificial intelligence systems are increasingly being applied to healthcare. In surgery, AI applications hold promise as tools to predict surgical outcomes, assess technical skills, or guide surgeons intraoperatively via computer vision. On the other hand, AI systems can also suffer from bias, compounding existing inequities in socioeconomic status, race, ethnicity, religion, gender, disability, or sexual orientation. Bias particularly impacts disadvantaged populations, which can be subject to algorithmic predictions that are less accurate or underestimate the need for care. Thus, strategies for detecting and mitigating bias are pivotal for creating AI technology that is generalizable and fair. Here, we discuss a recent study that developed a new strategy to mitigate bias in surgical AI systems.
More about this database...less...
Metadata for the creation of APA 7 citation:

Mittermaier, M., Raza, M. M., & Kvedar, J. C. (2023). Bias in AI-based models for medical applications: challenges and mitigation strategies. Npj Digital Medicine, 6(1). https://doi.org/10.1038/s41746-023-00858-z
Retraction Watch
Retraction Watch, acquired by Crossref in September 2023, is a database that lists retracted (retractions) or corrected (corrections) publications, or publications with an expressions of concern.
More about this database...less...
Metadata for creating APA 7 citation:

Retraction Watch. (n.d.). https://retractionwatch.com/

As a general rule of thumb, never upload any personal, patient, or healthcare information to a Gen AI tool.

In B.C., Collection, use, disclosure and retention of personal information is subject to and must comply with the provisions of the BC Freedom of Information and Protection of Privacy Act (FIPPA).

Balancing confidentiality and care coordination: challenges in patient privacy
Digital technology has significantly transformed healthcare, enhancing care coordination and improving patient outcomes. However, this transformation brings forth critical challenges, particularly in balancing the imperatives of confidentiality and efficient care coordination. Nurses play a pivotal role in balancing confidentiality and care coordination in the digital age, acting as guardians of patient privacy and critical facilitators of information sharing. Their unique position on the front lines of patient care requires them to navigate complex ethical and practical challenges.
More about this database...less...
Metadata for creating your APA 7 citation:

Ibrahim, A. M., Abdel-Aziz, H. R., Mohamed, H. A. H., Zaghamir, D. E. F., Wahba, N. M. I., Hassan, G. A., Shaban, M., EL-Nablaway, M., Aldughmi, O. N., & Aboelola, T. H. (2024). Balancing confidentiality and care coordination: challenges in patient privacy. BMC Nursing, 23(1), 1–14. https://doi.org/10.1186/s12912-024-02231-1

DSCI 321 - Health Analytics Library Research Guide

Generative Pre-trained Transformer (GPT)

The real skill you will need to use GenAI is knowing how to Prompt (explain) to the system what you want it to do.

This will take:

Time

To learn to write prompts that provide the information you need

You may have to run multiple iterations of your query.

Then, you will have to develop a research plan to identify and properly cite the information provided by the GenAI response.

You will need an above-average knowledge of medical and healthcare words and terminology to get proper value from using GenAI,

Money

The free versions of most GenAI applications are 'loss leaders' and are designed to interest you. For example, ChatGPT3 is free but limits you to the number of Prompts you can submit daily.

ChatGPT4 is a fee based service

The cost of using a GenAI applications total rest with the user

Using PICO(T) will help with writing better GenAI Prompts

GenAI hallucination is a phenomenon in which the large language model (LLM)—often a generative model—is inaccurate or made up.

The American Nurses Association (ANA) released a position statement on "The Ethical Use of Artificial Intelligence in Nursing Practice" in 2022. This report identified some critical considerations when working with AI.

REMEMBER GENAI HAS BEEN PRIMARILY TRAINED ON DATA FROM THE WEB

Bias to be aware of:

Under-representation of specific people or groups

Discrimination

Misinformation and Disinformation

Denialism

Plagiarization

Research Fraud

As a general rule of thumb, never upload any personal, patient, or healthcare information to a Gen AI tool.

In B.C., Collection, use, disclosure and retention of personal information is subject to and must comply with the provisions of the BC Freedom of Information and Protection of Privacy Act (FIPPA).