AI in Industry and Society by Davin Lam

Personalized Overview

davinlam — 2025-03-16 15:37:51 UTC

AI will affect every working industry that exists today. There is a common misconception that while industries such as transportation, finance, and software engineering are bound to be nearly entirely replaced by AI or automation in the near future, industries such as education, healthcare, and art are somewhat immune, owing to the importance of interpersonal relationships and creativity. The naivety of this belief has become increasingly obvious as new technologies within each industry emerge. Particularly in healthcare, it is unlikely that doctors will simply cease to exist. In many situations patients benefit greatly from interacting with and being reassured with who they perceive to be a caring and qualified provider. That being said, AI is already becoming a powerful tool in diagnostic imaging, patient monitoring, and the creation of personalized treatment plans. This is especially relevant to me, as a student with aspirations of working in physical therapy. AI tools have the ability to streamline and assist in practically every aspect of the treatment process short of directly speaking to and interacting with the patient. Understanding the ways in which AI can be leveraged in the industry will help me to be a better provider and take advantage of all possible opportunities to provide better and more efficient care. It is a simple inevitability that AI will be adopted into the industry. The only people at risk of replacement are those who are unable to adapt along with new technology and shifts in culture. This post will explore three of the most powerful tools in healthcare, discuss future trends and ethical considerations of AI in healthcare, and explore the broader societal impact of AI in healthcare.

Examples of AI Usage: IBM Watson Health

davinlam — 2025-03-16 15:42:38 UTC

IBM Watson Health, now known as Merative, is an AI driven tool that analyzes data in order to assist healthcare professionals in making informed decisions regarding disease diagnosis and treatment plan creation. The software assists providers by analyzing large quantities of literature, record data, and clinical trial data that would be otherwise impossible for a human, allowing for evidence based recommendations alongside human expertise. There are, however, clear challenges. Frameworks must be put in place to ensure that the software aligns with standard ethical care practices and maintains a degree of transparency so that providers are aware of the source of the advice that they may or may not incorporate into their care.

Examples of AI Usage: Tempus AI

davinlam — 2025-03-16 15:52:59 UTC

Similar to Watson Health, Tempus AI is an AI driven tool that analyzes clinical data in order to help healthcare providers make informed decisions based on a large pool of evidence. Using machine learning algorithms, Tempus has the ability to detect patterns in genomic data that can help predict a patient’s response to specific treatments. On the side of the patient, olivia, Tempus’ personal health app, helps patients better understand their own treatment by centralizing their health data. This allows patients themselves to make informed decisions along with their physicians. This, obviously, comes with serious data privacy concerns. Additionally, patients are especially susceptible to influence, making bias on the end of the algorithm significantly more dangerous, compared to a trained physician who can draw on their education to reject biased recommendations.

Examples of AI Usage: Medtronic

davinlam — 2025-03-16 15:57:26 UTC

Medtronic is a healthcare technology company that integrates AI into its hardware. For example, AI powered insulin pumps are able to detect, analyze, and predict glucose levels and modulate insulin release. This results in a more precise and individualized treatment of patients, on a scale that would be otherwise impossible. Such technology, obviously, comes with inevitable vulnerabilities and challenges. Safety measures must be put in place to ensure that the devices operate within safe parameters, and data breaches must be prevented in order to ensure the integrity of patient data privacy. Even so, patients can be more assured that even in the absence of a trusted provider, their care is being monitored and personalized.

Future Trends and Ethical Considerations

davinlam — 2025-03-16 15:58:38 UTC

The integration of AI in healthcare has already bolstered the efficiency of diagnostics, treatment creation, and chronic disease management. Going forward, these areas of patient care will only become more advanced. In particular, predictive analytics have become increasingly adopted and advanced. AI algorithms analyze patient data to predict disease onset and progression, enabling proactive interventions. Without concern for the limited manpower available in medicine, the algorithm can detect and act more quickly than any human provider. I am personally very aware of the reduction in treatment quality that can be brought about by the inability of a physician to attend to every single patient. It is an unfortunate reality that many people with chronic conditions must feel that they are being neglected simply because of the nature of the healthcare system. While an AI powered insulin pump or pacemaker isn’t a strict substitute, it can automate some of the process and leave many patients feeling slightly more at ease.

In a similar way, AI can also be used to analyze large amounts of a different type of data: genomic data. Such algorithms will be able to sift through and detect minute patterns in an inconceivably large pool of data. Naturally, the handling of genomic data comes with massive data privacy concerns and requires close regulation in order to be trusted. I am personally dubious of private companies having access to patient genomic data. Such a thing is too easily sold and exploited to insurance companies who are financially incentivized to have an upper hand in predicting patient health outcomes, even without their consent.

Societal Impact

davinlam — 2025-03-16 15:59:05 UTC

The adoption of AI in healthcare will have sweeping societal impacts in the realms of employment security, privacy concerns, and equity/bias. While I am not under the impression that swathes of the medical workforce will be wiped out by the advent of AI, the workforce will likely be restructured around AI. Just as how the digital EKG machine did not significantly reduce the number of hired cardiologists, but rather restructured their daily responsibilities, AI will likely do just the same. With many analytical roles being automated or streamlined, doctors will still be necessary to supervise decisions and serve as knowledgeable and empathetic overseers of a patient’s treatment, observing and taking into account a patient’s personal desires.

Privacy concerns are paramount in any scenario in which private firms have access to large amounts of patient health information. There is a serious ethical quandary involved in deciding whether or not private companies should be able to leverage personal health data with the primary goal of driving a profit.

Finally, it is of utmost importance to ensure equitable application of AI algorithms. Especially in the realm of genomic analysis, it is crucial that algorithms are not primarily trained and biased towards certain demographics of people. There have historically been problems associated with women and minorities being subject to treatment methods designed around research sampled from primarily white men. As a young, racial minority in the United States, I have often found myself having to speak to providers that are understanding of my situation in order to feel addressed and heard.

Reflection

davinlam — 2025-03-16 15:59:37 UTC

Prior to this research assignment, I was only vaguely aware of a few of the ways in which AI has been integrated into the healthcare industry. This assignment exposed me to technologies such as genomic analysis and AI driven insulin pumps, that genuinely surprised me with regard to their sophistication. While it was initially difficult to wrap my head around some of these, it ended up expanding my understanding of the future of AI in healthcare. More than before, I am confident that as AI is integrated further into healthcare, providers and patients may see greater levels of interaction, as analytical roles are replaced or streamlined by AI.

References

davinlam — 2025-03-16 16:00:15 UTC

IBM Newsroom. (2022, January 21). IBM announces agreement to sell healthcare data and analytics assets to Francisco Partners. Retrieved from https://newsroom.ibm.com/2022-01-21-IBM-Announces-Agreement-to-Sell-Healthcare-Data-and-Analytics-Assets-to-Francisco-Partners

Tempus. (2023). Tempus introduces 'olivia' - an AI-enabled personal health companion. Retrieved from https://www.tempus.com/press/tempus-introduces-olivia-ai-enabled-personal-health-companion/

Medtronic. (2025, February 25). FDA approves Medtronic's adaptive deep brain stimulation device for Parkinson's treatment. Retrieved from https://news.medtronic.com/2025-02-25-FDA-Approves-Medtronics-Adaptive-Deep-Brain-Stimulation-Device-for-Parkinsons-Treatment

The Guardian. (2025, January 13). Ministers consider allowing private firms to profit from NHS data in AI push. Retrieved from https://www.theguardian.com/society/2025/jan/13/ministers-mull-allowing-private-firms-to-make-profit-from-nhs-data-in-ai-push

AP News. (2024, December 9). Ethical challenges of AI in healthcare: A growing concern. Retrieved from https://apnews.com/article/231b4b7b8ed6a195ec920f1f362c15e2

The Australian. (2024, February 18). AI will pave the way for precision medicine, but risks loom large. Retrieved from https://www.theaustralian.com.au/health/ai-will-pave-the-way-for-precision-medicine-but-risks-loom-large/news-story/2ac35c2297efa378053246501844e81a

Barron’s. (2024, March 7). Tempus AI stock surges as AI-driven health tech gains momentum. Retrieved from https://www.barrons.com/articles/tempus-ai-stock-b1f2a144

Overview Video

davinlam — 2025-03-16 20:43:52 UTC

The assignment says to upload use the Padlet video recorder, which only allows 2 minute videos. The upload option also only allows a relatively small file size, much shorter than my 5 minute video. Thus, I have uploaded my video to the following Google Drive link. Let me know if this causes an issue with grading the assignment.