Whether you’re an avid user or avoid it as much as possible, artificial intelligence (AI) is here to stay.

When the telephone was new, some people resisted its use. They thought it had no practical use, that a telegraph was a better way of sending messages, and that it was dangerous and would intrude on privacy. [1]

It’s almost funny and quaint to think about these concerns now. The telephone changed the world. It brought opportunities for business growth and made people safer and more connected. Yet people sabotaged phone lines and refused to install them. [1]

New Technology Does Bring Valid Concerns

Some of the concerns were valid. Party lines (where more than one family shared a phone line) could mean that people listened to one another’s conversations. Telephone operators could also listen in after connecting two users. Telephone scams continue even today. [1]

People will always find a way to use a new technology for selfish or criminal purposes. But on the whole, you’d be hard pressed to find someone who thinks we should destroy all our landlines and go back to the telegraph.

We’re at the dawn of AI, which comes only 30 years after the internet became widely used, and about 20 years after social media was adopted on a large scale. The pace of technological advancement is relentless, so it’s no wonder that we find ourselves skeptical and rocked back on our heels as AI comes on the scene.

As a Content Creator: I Can’t Afford to Ignore AI

Much of the concern over AI comes from generative AI, which trains itself on massive amounts of data. That means, for instance, that AI crawls everything I’ve written here and on other publicly available platforms (such as Verywell and GoodRx, among others). It uses my writing to train, which means that if I go to an AI platform today and ask it to write an “IBD 101,” it’s likely using my words to create an article.

Creators are rarely asked whether their work can be used to train AI: the companies creating these models just did it. That means that people in creative disciplines have a real beef with generative AI, myself included. It can’t really “replace” us because it needs our work to train itself in the first place. But that hasn’t stopped companies from letting writers and other creatives go in favor of AI tools.

This leads some people to forgo AI altogether, which is a valid choice. However, it’s here to stay, so even those of us who are worried about losing our ability to earn an income will need to know how it is affecting us today and how it will affect us in the future.

AI Is in the Exam Room, Even If You Don’t See It

The use of AI in healthcare is exploding. Every day, I receive pitches and press releases about companies creating tools to leverage AI in hospitals, private practices, and academic institutions. There’s real difficulty keeping up with how AI is being used in healthcare because niche subsets of AI are evolving rapidly.

We do know a few ways that countries are turning their attention to AI. This includes [1]:

• 52%    Prioritizing areas for AI use in healthcare
• 64%    Diagnosing conditions (imaging and detection)
• 50%    Chatbots for interacting with and supporting patients
• 25%    Allocating funding for AI priorities

It’s clear that AI is useful in healthcare, and the reasons countries give for embracing and funding AI include[1]:

• 98%    Improving patient care (98%)
• 92%    Reducing workforce pressures
• 90%    Increasing efficiency and productivity

Some of the uses of AI in healthcare include [3]:

• Developing new medications and therapies
• Creating models to identify patients at risk for certain diseases and conditions
• Identifying babies at risk for congenital conditions before they’re even born

As patients, we may not even know how AI is being used behind the scenes in the management of our health. There are privacy and ethics concerns, because, for instance, our medical records could be used to train an AI. AI is also not free of bias, and it’s known to make mistakes because it’s only as good as the data it’s given. It has the potential to help all the patients that come after us, which is great. Still, we’re also justified in expressing concerns over privacy and in knowing how our data is being used.

In the IBD community, patients are across the spectrum in adopting AI or avoiding it. People who follow About IBD on Instagram told me they’re using AI for:

• Basic research about IBD
• Defining medical terms
• Explaining medical test results
• Taking notes during appointments

Some of the concerns community members expressed to me included:

• The unethical use of intellectual property to train generative AI
• Data breaches
• Privacy of personal information
• Energy and water use in AI data centers

One study on IBD patients used photos of stool (yes, literally photos of the toilet bowl) from people with acute ulcerative colitis. Patients then had a C-reactive protein (CRP) test. CRP is a blood test that’s an indicator of inflammation. The photos and the CRP level were compared and the AI was trained to understand, from the photo, which patients had a higher CRP level. [4]

This is not the only example of research into AI tools that are being developed for use in IBD. They won’t all make it to use by our GI teams, of course. Still, there’s a future where we can, within limits, have access to tools from home that tell us if we need updated or different care for our IBD.

We Can’t Expect Regulation, So We Need to Ask Questions

In the United States, federal regulation on AI has stalled. It is states that are turning towards passing laws to regulate the use of AI in healthcare. This will result in a patchwork of regulations across states, making it confusing and difficult for providers, payers, and patients alike.

AI has the potential to advance healthcare in ways we can’t yet conceive. But it also needs to be harnessed, and patients need to be part of the conversation about how our data is used and whether these AI systems are actually helping us. The open questions are numerous, including privacy, potential errors, and a lack of transparency (particularly from payers).

Maybe AI will be like the telephone, which put telegraphs out of business. The world moves on. But that doesn’t mean we shouldn’t be asking questions and demanding answers about how AI is already affecting us and how it will do so in the future. AI companies are benefiting from the lack of regulation and moving forward with products at incredible speed. Those two factors have the potential to leave patients and their privacy behind, which should concern all of us.

As AI tools become more sophisticated, patients must be part of the conversation and decision-making. Some of the things patients can ask of their healthcare team include:

• What types of AI tools are being used by your practice or hospital?
• Is my healthcare data being used for AI training?
• Will there be a disclosure statement regarding the use of my data by AI tools?
• If AI is used to interpret my lab or imaging results, will I be told? Can I request that it be reviewed by a human clinician as well?
• What are my options if AI makes a mistake in the delivery of my care?

These are difficult questions to ask and will likely also be challenging to answer. It’s unlikely we will see legislation or patient protections regarding AI in the next few years, so it’s important to become educated, stay engaged, and keep asking questions until we get answers.

Sources:

1. Elon University. 1870s – 1940s: Telephone.

2. World Health Organization. Is your doctor’s AI safe?

3. National Conference of State Legislatures. Artificial Intelligence and Health Care: A Primer.

4. Rotondo-Trivette S, Castelan VC, Mathur K, et al. Artificial Intelligence- and Physician-Interpreted Stool Image Characteristics Correlate With C-Reactive Protein Among Inpatients With Acute Severe Ulcerative Colitis: A Pilot Study. Crohns Colitis 360. 2024;6(3):otae043. doi:10.1093/crocol/otae043