EST 110 ZIHAO YUAN 116660660 S 6/7 - AI in Industry and Society by Zihao Yuan

Personalized Overview

zihaoyuan — 2025-10-12 23:06:47 UTC

My chosen topic will be AI within Healthcare. Healthcare is very dependent on the specialized expertise of professionals. This heavy reliance is now being supplemented by the use of AI. AI is now being used as a tool to support human capabilities, processing of huge amounts of data, offering diagnostic help, speed up development, and monitor patients. AI’s role in this industry is to offer itself to be an aid that lessens human workload but not replace it. I chose this topic because of my interest in healthcare. I often watch several health related videos on how the human body works. Most of these videos bring up potential future healthcare development. Sadly most of these ideas are very speculative and experimental, but with AI it might come into fruition. Within this Padlet we will discuss the implementation of AI in healthcare. We will go over 3 examples of AI use, such as image recognition and deep learning used for medical diagnosis, machine learning for analyzing large datasets for treatment, and patient monitoring tools powered by AI. Along with this, we will discuss AI trends, address its ethical considerations, and reflect on its societal impact. Healthcare is an important industry as it affects everyone. I believe AI will change healthcare greatly, changing it fundamentally. AI technology in healthcare can greatly increase health but it can also be easily misused. It can free up doctors but also be used to offer sub-par medical assistance. Especially working with something as important as human lives, AI implementation should be taken with great care.

Examples of AI Usage: AI in Medical Diagnosis

zihaoyuan — 2025-10-12 23:08:04 UTC

AI is being used for diagnosis. With the use of deep learning models and image recognition. AI can be used to analyze medical images such as X-ray, MRIs, and CT scans along with patient records. These models are trained using large amounts of data to identify patterns. By finding these patterns we can flag patients for signs of diseases. Once that is done a diagnosis can be given and treatment can be tailored. An example of this technology in use is IBM Watson for Oncology (WFO). WFO was the same Watson that won the 2011 TV show, Jeopardy. IBM decided to apply Watson to some real life applications. WFO analyzes patient records and then recommends treatment from a list, going from recommended, for consideration, and not recommended. An analysis of multiple studies found WFO to excel at identifying breast cancer with a 88.99% concordance rate with the Multidisciplinary Teams. Although this is great, the same analysis finds WFO struggles in other types of cancer such as the lowest, gastric cancer, with only a 57.94% concordance. A study with WFO in China suspects this as due to WFO only having access to only western data, the study conducted in China had a different demographic. This caused great discrepancy between the AI recommendation and local doctors. I believe that with better, more varied dataset that fit the demographics of patients we can implement this technology better. I think this technology can bring great automation to healthcare diagnosing, greatly reducing time. But anomalies must always be considered, and data must fit the patient to avoid AI bias.

AI in Medical Diagnosis Video

zihaoyuan — 2025-10-12 23:08:53 UTC

AI in personalized medicine Video

zihaoyuan — 2025-10-12 23:08:56 UTC

AI in patient monitoring Video

zihaoyuan — 2025-10-12 23:08:59 UTC

Example of AI Usage: AI in patient monitoring

zihaoyuan — 2025-10-12 23:09:01 UTC

AI is being used for monitoring patients. We can incorporate AI machine learning inside wearables and implants that can be used to track patient’s health. The AI can track a patient's heart rate, blood glucose, activity, and more. Any anomalies can be flagged and taken care of. This technology offers supervision for patients even outside of clinics. The few challenging issues are data reliability, and triggering false alarms. We also need to ensure algorithmic alerts are processed correctly. Personally, I think this can prove to be a great tool for monitoring. And like mentioned in a different example before, it can shift medical care to be more home based for better or for worse. But my main concern is the ethics and danger of constant monitoring of patients. Privacy should be made very secure to prevent this, but otherwise I see great potential.

Future Trends and Ethical Considerations:

zihaoyuan — 2025-10-12 23:09:04 UTC

I think AI implementation in healthcare will be more on a supportive side. As AI becomes more integrated it can analyze a patient's entire medical history, genetics, health activity, monitoring data, and other information. And with that to create treatment recommendations for doctors to approve. AI can be great tools for automation helping to relieve burdens on doctors. This tool has a lot of potential. But we also must take into consideration some ethical concerns. First we can start with algorithmic bias, if an AI is trained predominantly on data from one group, it will not work for others. To combat this we will need to ensure varied reliable data. Secondly, we have the AI black box issue. AI reasoning is hard to understand and because of that we can never fully trust it. To solve this problem we need to find a way to make AI more transparent, such as explaining its reasoning. Thirdly, we have the issue of privacy and cybersecurity. AI requires vast sensitive amounts of health data. All of this sensitive information must be stored. The ethical concern is should companies have this data? And the danger of cybersecurity breaches of these data can cause. Especially when it is about someone’s physical health which cannot be changed. The final ethical issue is, who is to blame when the AI fails? My personal insight is that AI will develop to be mostly able to address these issues. What AI implementation needs is better data, transparency, strong cybersecurity, and always an option to opt out of AI use with no downside.

Societal Impact:

zihaoyuan — 2025-10-12 23:09:06 UTC

I think AI inside healthcare will have a great impact. It can offer a lot of good but also bad. AI automation already reshapes industries it is present, replacing old jobs but offering new jobs like data management and AI supervision. AI automation can also let doctors focus on their more specialized tasks instead. But for specially patients, we can see AI bringing medical help closer to home, which is great if this can be offered to all. But if not it can further societal divide. AI use will also mean patients will have to offer up a lot of their privacy for the service. So much sensitive information being available becomes really big targets for cyber attacks. And even when not being stolen, such data can be used for discrimination. Important data like those can be used by employers or insurers to change how they treat someone. If AI in healthcare becomes prevalent we must have robust cybersecurity, and clear privacy protection, and options to not use AI. Personally, I believe AI will always remain a tool. AI cannot be relied on. AI can enhance healthcare, and if implemented well can make healthcare more accessible. But AI development must work with doctors instead of replacing them. AI implementation must also be done carefully to ensure no misstep, which in cases like healthcare can be devastating to individuals. Policy makers have to watch AI closely to enforce this. The end goal of all of this should be to make people healthier in the end.

Reflection:

zihaoyuan — 2025-10-12 23:09:07 UTC

During this research process I have encountered many ways AI has been used within the healthcare industry. I have informed myself of all its upside and its downside. AI will not be a direct upgrade to what we have before. Instead it will build up and enhance what we currently have. During research a lot of common themes about privacy have been brought up. Indeed AI uses a lot of data to do its job which is sometimes harmless. But the data in healthcare is very sensitive and personal, and misuse of it is a great privacy risk. The challenge while conducting research has to be finding good sources. AI discussion is a bit sparse especially for very specific topics. Research studies are especially hard to find. In the end I just had to keep on digging for sources to find what I want. I think AI will greatly benefit healthcare. If it is done with care it will make everything so much more accessible. Not only will it be easy to access AI, AI can free up human resources so human aid can be easy to get too. But the risks cannot be ignored too hence the caution. Personally, this shaped my thoughts on AI to be more cautious. I have seen all of its innovations and potential, but I am also now aware of its drawbacks. If it is not done proper the damage can be great.

Example of AI Usage: AI in personalized medicine

zihaoyuan — 2025-10-12 23:15:14 UTC

AI is being used for personalized treatment, using AI to analyze different types of data such as genomes, clinical records, and lifestyle habits. Machine learning models can be used to identify correlations between genetics, disease progression, and drug responses. Such a case is Tempus. It analyzes a patient's clinical data, with molecular data, and a bunch of other things. Then it returns test results of what it found. WIth this technology massive amounts of different types of data can be analyzed and a result can be made. The few challenges facing this AI is algorithmic bias and how to combat wrong diagnosis. Sometimes the wrong things can be prescribed and if it doesn’t have an immediate effect, the lack of giving the proper medication can be dangerous enough. The AI should be trained well to avoid such cases and have human supervision. I think this can be significant in that it changes how in the future we will get medical aid. The ease of AI can shift people to receive aid more at home, which can be great for convenience but also isn’t as great as standard hospital visits. So personally I think AI should be relegated to the side and hospital visits should still be the standard to avoid any mis-diagonising.

References:

zihaoyuan — 2025-10-12 23:15:21 UTC

Jie, Z., Zhiying, Z. & Li, L. A meta-analysis of Watson for Oncology in clinical application. Sci Rep 11, 5792 (2021). https://doi.org/10.1038/s41598-021-84973-5
- provided an analysis of multiple studies on IBM Watson for Oncology, responsible for the
Henrico Dolfing. Case Study 20: The $4 Billion AI Failure of IBM Watson for Oncology. December 07, 2024.
https://www.henricodolfing.com/2024/12/case-study-ibm-watson-for-oncology-failure.html
- provided background information about Watson for Oncology and why its shortcomings
Zhou, N., Zhang, C. T., Lv, H. Y., Hao, C. X., Li, T. J., Zhu, J. J., Zhu, H., Jiang, M., Liu, K. W., Hou, H. L., Liu, D., Li, A. Q., Zhang, G. Q., Tian, Z. B., & Zhang, X. C. (2019). Concordance Study Between IBM Watson for Oncology and Clinical Practice for Patients with Cancer in China. The oncologist, 24(6), 812–819. https://doi.org/10.1634/theoncologist.2018-0255
- Provided the specific study performed with IBM on patients in china, attributed poor concordance rate with training data conflict with region.
https://www.tempus.com/about-us/tempus-tech/
- The Tempus website itself, supplies helpful background info.
Al Gyani. Tempus AI – Precision Medicine Platform
https://aigyani.com/tempus-ai/
- Analysis on Tempus, provided info on how it worked, along with pros and cons.
https://www.medtronic.com/en-us/index.html
- the Medtronic website itself, supplied helpful background info
Sapp, J. A., Gillis, A. M., AbdelWahab, A., Nault, I., Nery, P. B., Healey, J. S., Raj, S. R., Lockwood, E., Sterns, L. D., Sears, S. F., Wells, G. A., Yee, R., Philippon, F., Tang, A., & Parkash, R. (2021). Remote-only monitoring for patients with cardiac implantable electronic devices: a before-and-after pilot study. CMAJ open, 9(1), E53–E61. https://doi.org/10.9778/cmajo.20200041
- study used to see medtronic monitoring system effectiveness, study found it reduced costs
Pham T. (2025). Ethical and legal considerations in healthcare AI: innovation and policy for safe and fair use. Royal Society open science, 12(5), 241873. https://doi.org/10.1098/rsos.241873
- AI in Healthcare, goes over ethics

-

zihaoyuan — 2025-10-12 23:32:21 UTC

Overview Video

zihaoyuan — 2025-10-12 23:32:24 UTC

Padlet had a 2 minute limit, had to use clipchamp from previous assignments instead.

-

zihaoyuan — 2025-10-12 23:32:26 UTC

-

zihaoyuan — 2025-10-12 23:32:28 UTC