AI could diagnose your illnesses

family checkup doctor

No more sticking out your tongue and saying “ah”; snapping a selfie and sending it off to a website or submitting it through an app might be enough for a medical diagnosis.

Artificial intelligence is starting to hit its stride in the healthcare industry. Instead of worrying that doctors and nurses will be replaced by robots or, worse, eliminated altogether, patients will likely benefit from deep learning, better analytics, more specific treatments and new procedures that could not only cure what ails them today but protect against more severe conditions tomorrow.

Consider devices like FitBits and other wearable technology and apps that can track health indicators like pulse, blood pressure and activity rate. People think nothing of wearing something akin to a watch that reminds them to stand up, drink water, go for a run, etc. It is highly plausible that as this technology expands and becomes even more commonplace, so, too, will users become more comfortable allowing those devices to provide their data (with permission) to doctors and researchers for crowdsourced information. It could provide the kind of broad and deep data that will allow for larger sample sizes for trials or statistical analysis to make better health determinations and predictions for small communities and larger populations.

Market Map from “From Virtual Nurses To Drug Discovery: 106 Artificial Intelligence Startups In Healthcare” (CB Insights, 2017)

There’s also much to be gained from improved data mining of existing health records and self-provided information. Forbes anticipates AI data mining could be a $6 billion market by 2021, and Amazon has already discussed plans to sell software to hospitals and healthcare facilities to help mine data for the sake of improving health records and treatment options while reducing overall costs.

The idea of robots as surgeons is one thing human patients have come to embrace, knowing that robots are controlled by highly skilled experts while capable of making incredibly small, precise incisions, leading to shorter recovery time. But eventually, even the robotic surgeon could be operated by AI. Currently in the early stages of its training, the Smart Tissue Autonomous Robot can make smaller, cleaner sutures than a robot guided by a person.

It’s critical that in health tech, as it should be across the development of all technology, that people come first.

AI and physicians can work hand-in-hand, so to speak, to make better clinical decisions and more accurate diagnoses. AI might replace humans in some small tasks, by relying on deeply mined data to draw from research and recognize potential health issues that could be covered by other symptoms, but it will not replace humans overall.

Healthcare overhaul

Doctors, indeed, could be capable of serving patients in remote areas in which their expertise is not otherwise available. That’s one of the 12 ways in which AI might disrupt the healthcare industry, according to the 2018 World Medical Innovation Forum on AI.

The “Disruptive Dozen” for healthcare include, in part:

  • Brain-computer interfaces powered by AI that could restore neurological activity to patients who have suffered severe trauma, allowing them to regain lost abilities including speech and movement;
  • Radiology tools that allow for even more in-depth images of the body’s internal workings without requiring biopsies and which could allow for a better understanding of how tumors behave and better identification of aggressive types of cancer;
  • Imaging tools that could provide ultrasound and radiological exams remotely to diagnose illnesses like tuberculosis in areas where highly specially trained technicians are scarce;
  • Using natural language processing tools for more accurate voice recognition and dictation, cutting back the amount of time doctors and nurses need to spend transcribing their notes and allowing them to spend more time on patient care;
  • Better interfacing and analysis of electronic health records to identify infection patterns and indicate patients that could be at risk before an infection sets in, which could help to combat the spread of antibiotic resistance;
  • Creation of more precise analytical procedures for reviewing pathology results, getting down to the pixel level to determine a more exact course of treatment;
  • Make smart devices even smarter, with better tools to find and alert medical staff to minute changes in a patient’s condition for more prompt treatment;
  • Taking the emerging world of immunotherapy and using machine-learning algorithms to gain better insight into disease biology and develop better treatments;
  • Use of electronic health records as a risk indicator and prediction tools not just for future procedures but for indications of a growing problem or factors leading up to a possible health event like a stroke or heart attack; and
  • Enlisting predictive analytics to determine the best course of action not for symptoms on display today but for what the person’s overall health says about their future risks.

It’s critical that in health tech, as it should be across the development of all technology, that people come first. When it comes to personal data, especially health records, we’re reluctant to give access to large institutions. But if allowing access to our data means improving healthcare and living conditions for all, being able to predict and therefore prevent serious health concerns, and create safer procedures, then I think we can all work together to create a better future.

About Fast Future

It is our mission to explore the implications of emerging technologies, seeking answers to next-level questions about how they will affect society, business, politics and the environment of tomorrow.

We aim to inform and inspire through thoughtful research, responsible reporting, and clear, unbiased writing, and to create a platform for a diverse group of innovators to bring multiple perspectives.

Fast Future is building the media that connects the conversation.

More
cute robot illustration
Designing loveable robots