Within the first five days of being launched, ChatGPT had over 1 million users. However, ChaptGPT’s advances have created polarizing reactions: While many point out how AI is more accurate than humans, can analyze large sets of data, and offers 24/7 availability, others note its repercussions. It could result in 45 million people losing their jobs, a loss of transparency, and ethical dilemmas. Yet instead of viewing the effects of AI in these totalizing ways, it is crucial to recognize its different impacts on different industries. Looking at three key industries that AI will transform—healthcare, the legal system, and security—the effects of AI range from beneficial to harmful depending on particular circumstances.

On one end of the spectrum, the effects of AI in healthcare are undeniably advantageous. A 2016 Stanford study showed that AI could diagnose lung cancer with greater accuracy than a human pathologist while cutting medical costs and wait times for patients. As AI continues to develop, robotic surgeries will also become widespread—increasing success rates of various medical procedures and increasing patient accessibility. In fact, a variety of robotic surgeries have already been performed. These surgeries allow for smaller incisions, reduced blood loss, quicker healing, and less pain than standard human-led operations. The question should not be if AI should be implemented, but to what extent should AI be implemented. 

Yet in other fields, such as the American legal system, the case for AI intervention is not so clear-cut. With AI’s ability to process large amounts of data and store it in its unlimited memory, it can process hundreds of cases in parallel while exploring all of the cases’ details—something human judges cannot accomplish. Even though the benefits of robojudges are compelling, the adverse effects are equally so. AI models are built using neural networks, a complex system of mathematical computations to produce a single output: in this case, guilt or innocence. As a result, the only explanation to why a defendant was convicted  is that the AI trained on large amounts of data and simply made a decision. While robojudges have their advantages, there is a tradeoff between unbiased efficiency and the loss of personal freedom and transparency. Unlike AI in healthcare where AI can save lives, this is a question of personal preference where there is no right or wrong. 

Finally, in sectors such as security, the harms of AI greatly outweigh its benefits. Even though AI can improve security by preventing leaks and quickly detecting software bugs, AI can be used by hackers to find new vulnerabilities in a system and perform intricate hacks. In 2014, when a Generative AI model, GANS, was introduced, the AI could produce various types of content from videos to audio recordings. In 2019, PassGAN, an AI model developed using GANS, cracked 51 percent of user passwords in under a minute. It’s already proven that AI can be used maliciously for hacking users’ computers, and as AI continues to evolve, these problems will only worsen. 

When the effects of AI are analyzed with respect to specific industries such as healthcare, law, and security, the extent of the benefits and risks of AI are drastically different. While AI’s benefits in healthcare seem to far outweigh possible downsides, there are more ambiguous cases such as the role of AI in law. Lastly, the implementation of AI in security favors the perpetrators with new developments in hacking and phishing that may prove disastrous. The key point is that the effects of AI must not be analyzed universally, but through particular cases within each industry. 

References:

Tegmark, M. (2018). Life 3.0. Penguin Books.

Kubota, T. (2017, November). Algorithm better at diagnosing pneumonia than

     radiologists. Stanford Medicine. med.stanford.edu/news/all-news/

     2017/11/algorithm-can-diagnose-pneumonia-better-than-radiologists.html

Silajdzic, M. (n.d.). AI can crack 51% of passwords in less than a minute:

     Study. VpnOverview. vpnoverview.com/news/

     ai-can-crack-51-of-passwords-in-less-than-a-minute-study/

The pros and cons of artificial intelligence. (2021). Forbes.

     forbes.com/sites/qai/2022/12/01/

     the-pros-and-cons-of-artificial-intelligence/?sh=524e193d4703

Pros and cons of artificial intelligence. (n.d.). Intellipaat.

     intellipaat.com/blog/pros-and-cons-of-ai/