I’ve always been a fan of sci-fi movies. Classics like Blade Runner, starring Harrison Ford, Alien, featuring Sigourney Weaver, and The Terminator with Arnold Schwarzenegger, were favorites. One of Schwarzenegger's lines that always stuck with me was “Talk to the hand." It often comes to mind when I find myself battling unhelpful automated customer service systems. There’s something deeply frustrating about trying to communicate with a machine that doesn’t truly understand you; it really does feel like talking to the hand.
What’s remarkable is how those decades-old films seemed to anticipate the rise of machines and artificial intelligence (AI). It raises an interesting question: is life imitating art, or were those films simply ahead of their time?
One scene from Blade Runner exemplifies this question: Ford’s character zooms in on a photograph using voice commands to identify a synthetic snake used for the show by a humanoid killer. At the time, that level of technology felt like pure science fiction. Yet now, we’re seeing similar capabilities emerging – particularly with image analysis tools for digital pathology slides.
Digitizing pathology has long been both a dream and a professional mission of mine. Throughout my career at various academic institutions, I’ve worked to help communities transition their pathology practices into the digital age, though not always with immediate success. Today, as a bone pathologist at the Hospital for Special Surgery, I signed out all my cases digitally. It’s been a transformative shift for my workflow – more efficient, better organized, and far timelier.
While the field of pathology is some way off its complete digital transformation, AI is already having a significant impact, particularly in breast and prostate diagnostics. AI tools capable of scanning and analyzing vast numbers of slides can streamline workflows and flag potentially suspicious areas that might be missed due to high caseloads.
Pathologists deal with immense diagnostic pressure every day. With such a wide variety of specimens, the risk of overlooking subtle but crucial findings is real. AI can serve as a second set of eyes, helping pathologists work both more efficiently and accurately.
Used thoughtfully, AI can help accelerate progress. But if we lose control of the tool, do we risk being controlled by it?
Personally, I see AI as a powerful tool – one that, if used wisely, has the potential to advance society in meaningful ways. But humans must remain in control. We must never allow ourselves to become subordinate to the very technologies we create. AI development must always remain under human oversight. As per a quote I came across: “I want AI to work for me – not work me out.”
The goal isn't to let technology replace us but to ensure it serves us. That balance is essential for the future. Some software can suggest differential diagnoses based on set parameters, but pathology is rarely straightforward. Distinguishing between reactive and malignant processes often involves subtle nuance decisions that require human expertise and directly affect patient outcomes.
From my perspective, AI is already making a difference in pathology, and its role will only continue to grow. Since pathology is foundational to accurate diagnosis, AI can enhance our ability to deliver more precise and timely results.
Integrating AI into pathology workflows has the potential to enhance patient care, minimize diagnostic errors, optimize healthcare resources, and streamline processes. While I’m not yet at the point of incorporating AI into my daily diagnostic routine, I firmly believe that – with care and oversight – it can be a valuable partner.
Ultimately, AI is most effective when it complements, rather than replaces, our medical judgment. Relying entirely on AI for diagnostic decisions is risky and must be approached with caution.
