Sherwood, A. 2023. How will generative AI impact healthcare?. (online) Geneva: World Economic Forum. Available from: https://www.weforum.org/agenda/2023/05/how-will-generative-ai-impact-healthcare/

(Accessed: 10/11/2023)

Aaron Sherwood is a current genAI entrepreneur in the technology industry. His background as a founder of Global Shapers a community for under 30's working to tackle local and regional global challenges. World Economic Forum is an independent international global organisation with an aim to influence and policy to benefit society. Its readership stretches across the world.

The author makes light of the challenges faced by the healthcare industry, including; pandemics, aging populations and rising costs. He argue it is critical to address the potential offered by genAI tools to combat these challenges. The author is positive towards genAI but does advise the need for quality datasets. The author provides a rational and balanced argument and offers a perspective from the technology sector.

A study was conducted to assess the viability of using genAI to communicate health information with patients. The study used a survey approach. Patients provided questions, 5 responses were AI generated and 5 from a healthcare provider. Patients were asked to distinguish between human and AI. The study found there were difficulties in differentiating between human AI. Generally, the information produced by AI was accurate. Feedback from participants on trust showed low levels for complex medical queries. The authors concluded by saying there is potential for genAI to be used for communicating medical information, but further research is needed.

The study has been cited 7 times according to PubMed and the publishing journal has a H rating of 23. The authors are colleagues at New York University.

The study provides a scenario of using genAI in a more risky area, with more ethical risks and implications.

The Report assess the effectiveness of ChatGPT in healthcare by looking at its abilities to draw conclusions and comparing its data analysis with actual data collected. The outcome, potential benefits were identified such as improving scientific literacy, extraction of information from medical texts, simplifying complex research into easily readable language, and handling complex data loads. The study noted concerns with ethics and limitations from bias and hallucinations and emphasized the importance of being educated in using genAI appropriately and an awareness of its limitations.

This article has been cited 72 times and is rated as a 'highly cited paper' on Web of Science. showing its influence in this area. The journal has a current H rating of 100.

It has been included for its influence and its argument for education on using AI which is an import factor in information literacy skills.

Avaliable at: http://dio.org/10.1002/ohn.465

Authors carried out a study to compare Google search engine and ChatGPT's medical knowledge capabilities. They were tested by asking both a series of questions from the perspective of the patient. The study found ChatGPT demonstrated more accurate general medical knowledge, but Google scored higher for health recommendations. Authors suggest medical staff should try to gain an understanding and awareness of these ramifications so they can appropriately advise their patients.

The publishing journal has a H-Index rating of 136 and is part of Sage publications Inc, a prominent social science publisher. The article has been cited 2 times. The authors are both members of Stanford University Medical school, one of the leading schools for biomedical research.

As ChatGPT seems like a natural successor to web engines and both arguably, are used for quick medical advice, the comparison is important.

The author advocates for using a devised framework to optimize the use of large language models (LLMs). Five principals make up their CLEAR framework. This stands for: Concise, Logical, Explicit, Adaptive and Reflective. These act as guidelines to assist the user to evaluate and create high quality content with generative AI tools. The author argues for librarians familiarizing themselves with the framework and incorporating it into information literacy skills set.

This was taken from a journal with a H-Index rating of 64. The article has been cited 11 times since publication. The author is a professor of College of University Libraries and Learning Services at the University of New Mexico. He has been cited 211 times.

The article was chosen to reflect the possible impact on information literacy skills and, a possible solution to harnessing genAI tools to their full potential.

The authors suggest doctors and health professionals need to apply caution when using ChatGPT and similar genAI tools due to lack of authentication of results or peer review. Subsequently, in extreme circumstances, health professionals may find themselves facing medical malpractice cases. They recommend using genAI tools to compliment other reputable search techniques and sources and to compare answers to check for accuracy.

JAMA forum is an online, peer-reviewed, open access journal. It is part of the popular JAMA network which has a high impact factor of 13.8 for journals in the health science sector. The article has been cited 8 times. The primary author is part of Stanford University's medical and law schools, suggesting a high knowledge of both topics.

The article raises legitimate concerns for health professionals and the possible implication of misuse of genAI. It could deter medical staff from using genAI tools.

An editorial piece taken from Molecular Therapy-Nucleci Acid which has a H-index rating of 84 and impact factor of 8.8. It is part of Cell Press who produce scientific content. Primary author Charkraborty has been cited over 5000 times and a individual H rating of 43.

The authors present a positive case for the use of large language models (LLMs) in the drug discovery and development process. The drug development process has historically been a lengthy and costly procedure. Companies have adopted new business models based on AI, deep learning and quantum mechanical method. This has proved beneficial in reducing cost and time. LLM's are also being used to help with explaining drug interaction and drug re-purposing.

This piece provides an example of genAI being used effectively in an area of healthcare. It highlights the cost and time saving benefits that are often mentioned.

This peer reviewed article provides a summary of review and research on large language model (LLMs) use in healthcare. How LLMs can be utilized and potentially benefit areas in healthcare is mentioned. Authors suggest they can be used to improve efficiency in clinical services, extracting information from health records and using research. Concerns are made about the potential for dangerous and offensive responses from the LLM, and possible data breaches. Another concern made by the authors is the conflict with "the right to be forgotten" if personal data is incorporated during model training. Authors argue for more rigorous research.

The journal is a prominent publication in the field of medicine with a H-index rating of 605 and impact factor of 82.9. The article has 28 CrossRef citations proving its influence.

It was chosen for its balanced view and consolidation of research in the area.

This is a clinical review and education piece about how large language models (LLMs) can be utilized in medicine. Authors state LLMs are being used to perform medical related tasks without the LLM receiving specialist training and input on medical data. They argue for more involvement from medical professionals in shaping and evaluating these tools in real-world testing with the goal of the LLM to expand human judgement rather than replace.

JAMA is a widely circulated journal group with a H-index rating of 739. It has a large readership with 27 million annual hits to their website. Dr Shah is Proffessor of Medicine and Chief Data Scientist at Stanford University. He has numerous awards in bioinformatics. The article has been viewed 59,000 times.

The stance presented is noteworthy as it encourages active participation from medical professionals and highlights the current lack of training/support on LLMs.

(Accessed: 02/11/2023)

This is a summary of a legislative proposal for an Artificial Intelligence Act recommended by the European Commission. The aim of the act is to govern the use of AI systems within the EU, assign a technology-neutral definition of AI systems and ensure these systems comply with EU values and principles. Systems will be classified by risk, and practice that causes harm will be prohibited. Systems deemed high risk will require greater monitoring.

The author is a policy analyst for European Parliament and has a PhD in competition law and telecoms regulations with an interest in regulatory affairs and European law, proving his credentials as a reputable source.

The proposed act is the first of its kind. If passed it will affect how users can interact with genAI and its influence will be felt globally. Most research notes the lack of regulations and ethical guidance around genAI as a major challenge.

Pictures were created by typing in prompts and descriptions related to the citation report.