Skip to Content
Menu

AI in Healthcare: Potential Uses and Drawbacks


Artificial intelligence (AI) is a tool. “Any tool can be used for good or bad,” the visual effects artist John Knoll famously said. In the healthcare space, AI has the potential to make medicine more personalized and accurate. But concerns exist about the best ways to use AI responsibly.

According to a Pew Research survey, a majority of the American public is not convinced AI in healthcare can improve outcomes. This doubt may not be justified in every case. But the facts support having a healthy amount of skepticism.

Learning more about the state of AI in healthcare can help empower you to recognize its potential uses and drawbacks.

What Do People Mean When They Talk About AI in Healthcare?

Interest in and adoption of AI has seen a significant increase since 2023. Broadly speaking, AI is any machine used to mimic human intelligence. Different types exist. In the past few years, when people refer to AI, they often mean a type called generative AI.

Generative AI uses large amounts of data to make text, pictures and other content. Patients may use generative AI in certain situations, like researching how doctors diagnose cancer.

For example, you may sometimes get an AI result when you search Google. Google’s generative AI creates these responses by drawing from data used to train it. ChatGPT is another example of a generative AI device.

This image shows an example of AI-generated text answering the question "how is mesothelioma diagnosed?"

Doctors may use generative AI for certain purposes, like reducing documentation time. They may also use other AIs, including:

  • Computer vision: This type of AI can interpret and analyze images. For example, AI computer vision may be able to help detect lung irregularities in CT scans.
  • Robotics: This type of AI performs routine, repetitive tasks. For example, AI-equipped robots exist to help surgeons monitor vitals during surgery.

Experts continue to develop new applications for AI in healthcare.

How Powerful Is AI Today?

The most powerful AI that exists today is called narrow or “weak.” Despite the names, it still has many useful applications. But it can only perform a single or narrow task. For example, generative AI can only produce text or images.

General or “strong” AI is popular in science fiction, but it is still just that. General AI is a theory that describes an AI that can perform any intellectual task a human can.

What Are Some of the Major Concerns About AI in Healthcare?

Increased interest in AI has caused concern among doctors and patients alike. The AI field is evolving quickly. Many questions exist about how to use AI tools responsibly. Major concerns about using AI in a healthcare setting include:

  • Hallucinations: AI sometimes “hallucinates.” This describes when AI produces incorrect or misleading answers to questions.
  • Loss of the human element: AI, used irresponsibly, could negatively affect doctor-patient relationships. Patients in one survey worried about the implications of AI creating more impersonal healthcare.
  • Risk of bias: AI may reflect biases in data collection. For example, racial biases in data sets may negatively affect the predictive power of the tool.

Generative AI and Mesothelioma Research

Generative AI tools like Google’s AI search results are only as good as the data used to train them. Because mesothelioma is such a rare cancer, the risk for bad search results may be higher than for other cancers. Patients and families may want to find more reliable sources and talk directly to an expert.

With these concerns in mind, it is important to be aware of what AI can and cannot do. In some cases, it can be a very useful tool. But right now, it has serious limitations that should be taken into consideration.

What AI Can and Cannot Do for Patients and Doctors

AI may be able to:
  • Help doctors decrease administrative tasks that limit personalized time with patients
  • Help doctors diagnose diseases using computer vision to analyze imaging scans
  • Help patients find products and services online
  • Help patients learn more about their condition, like explaining medical terms or how cancer is diagnosed
AI is unable to:
  • Provide a diagnosis on its own without human input and analysis
  • Provide advice to patients, like deciding whether to go to an emergency room for chest pain
  • Provide up-to-date medical information, as AIs may rely on info that has not been updated in a long time
  • Replace interactions with healthcare providers

Tips for Patients Using AI for Healthcare Purposes

Patients may want to maintain a healthy dose of skepticism when it comes to using AI for healthcare purposes. Generative and other types of AI are still in the early stages of development. Questions remain about their accuracy, equity and applications.

When using generative AI for medical purposes, patients may want to:

  • Check the sources: Some AI tools will show you the sources of the information provided. This information can help give you insight into whether the AI tool is generating trustworthy information. But double check before you decide if you can trust it. Sometimes, the tools provide references that do not verify the AI’s statements.
  • Limit uses to basic education and resource searches: Generative AI can be a useful educational tool because it can synthesize a lot of information. For the same reason, it can help compile resources. Beyond this, its uses for patients are limited. And even these two uses are still subject to problems like hallucinations and bias.

Patients should never be afraid to reach out to a medical expert – you can even ask them for their thoughts on AI in healthcare.

AI is not meant to replace personalized medical advice and treatment, nor is it capable of doing so. For mesothelioma patients, doctors specializing in this disease practice at cancer centers throughout the country. These experts have perspectives that cannot be found in any known AI tools.