Greg Freiherr, Industry Consultant

Greg Freiherr has reported on developments in radiology since 1983. He runs the consulting service, The Freiherr Group.

Blog | Greg Freiherr, Industry Consultant | Artificial Intelligence| July 05, 2017

Is There A Centaur In Your Future?

artificial intelligence

I miss carbon paper. I get that it was messy. The imprint of characters was incomplete, no matter how hard I struck the keys. But its presence between two sheets of paper — original and carbon copy — was ... comforting. Physically putting it there and scrolling the carbon sandwich into place would get me something.

It was a mechanical exertion easily within my ability. The result may have been minimally effective. But it was effective.

Alas, carbon paper is gone. As is the manual typewriter. And white out. (I loved white out.) Gone too are the electric typewriter and its backspace, correcting cartridge.

They are not the only casualties of progress: my beloved public phone booth; the answering machine; the home phone; the phone itself. (What remains is a wireless, portable electronic communications device for creating, sending and receiving texts and pictures. Nobody uses the “phone” to talk anymore — unless the talking is accompanied by video (e.g., Skype, Facetime).

What will go next? And how will its fall preceded others as have the domino-like staples of my once modern life?

 

Where You Won’t Go

Dr. Seuss used to speak to my imagination. So thrilled was I by the enduring concept of Seuss-like exploration that I took “Oh, The Places You’ll Go!” out of deep storage to show one of my sons on his college graduation. It was comforting to think Seuss was still relevant! But was he? Is he?

Machines are exploring the universe. They have been for a long time. Viking, Voyager, Cassini. Neil Armstrong and Buzz Aldrin were aberrations of the Cold War.

From algorithms that prioritize scans to ones that offer expert advice about interpretation, artificial intelligence is coming to radiology. Does that mean smart machines coming for radiologists? Will they take over? Will they make radiologists obsolete?

The worry that machines will replace humans is very old hat. Twenty years ago, on May 11, 1997, IBM’s Deep Blue defeated chess grandmaster Garry Kasparov. But human chess masters didn’t go away. In fact, you could argue that the opposite has happened.

Computer programs for Apple and Windows operating systems have “taught” people to play the game. Newbies who might not have otherwise joined in are playing the game. Computer sparring provides practice. People have gotten better.

Computers have even “evolved” human play through the ne’er before creation of computer-human chess teams, sometimes referred to as a “centaur chess player.” Could a similar future be in store for radiologists?

 

Gaming The Answer

Chess was once thought to require human intelligence. This is why developers of AI targeted this game. But, since Deep Blue’s victory, chess has been replaced by Go as the common denominator of what some modern day disciples of Alan Turing believe make people unique. Not surprisingly, last year, Google’s AI game-player, called AlphaGo, defeated a human champion of Go, which was invented in China some 3,000 years ago. Don’t be shocked. Or dismayed.

The problem is not that computers are eclipsing humans. It is that we are confusing the signs of
intelligence with intelligence itself — the savant
with the genius.

The ability of machines to master specific tasks and do them in savant-like fashion underscores the unsung value of AI for humanity: the ability of AI machines to master specific tasks through unrelenting effort. Their drive to succeed comes not from ambition but from human overseers pushing soft-wired buttons.

AI is like a dog trying to rip a play toy from the hand of its master. Its focus is, shall we say, limited. As such, it could be a godsend to radiology.

AI could be the strong, if not better half, of a computer-human pairing that encourages physicians to take up radiology as a specialty, accelerates their learning, then serves as a dedicated — and very aloof and non-threatening — sidekick that raises human performance through expert advice made available for the interpretation of routine cases.

Maybe my belief in such a future is rooted in the history of mankind to evolve. Maybe it’s an emotional repudiation of the singularity, which some of the great thinkers of our time (e.g., Hawking, Musk, Gates) say will mark the demise of humanity.

Maybe it’s just my deeply engrained love of carbon-based life.

 

Related Content

RSNA 2017 Celebrates Innovation in Radiology
News | Imaging | January 15, 2018
January 15, 2018 — The Radiological Society of North America’s...
Transpara Deep Learning Software Matches Experienced Radiologists in Mammogram Reading
News | Computer-Aided Detection Software | January 12, 2018
Deep learning and artificial intelligence improves the efficiency and accuracy of reading mammograms, according to...
The artificial intelligence (AI) smart algorithm onboard the Infervision stroke product calculates the volume of bleed on the basis of multiple brain CT slices.

The AI smart algorithm onboard the Infervision stroke product calculates the volume of bleed on the basis of multiple brain CT slices. The size of the bleed volume indicates the urgency and type of treatment required.

Feature | Artificial Intelligence | January 10, 2018 | Greg Freiherr
Exhibitors at the 2017 Radiological Society of North American (RSNA) meeting rode the...
Artificial intelligence was the number one topic in radiology in 2017.

Artificial intelligence was the number one topic of interest in radiology in 2017, based on the most popular articles and videos on ITN in 2017.

Feature | Imaging | December 28, 2017
The Imaging Technology News (ITN) website had another record year with more than 1.25 million page views in 2017.
Median Technologies Collaborating With Chinese Hospital for Lung Cancer Screening Programs
News | Lung Cancer | December 22, 2017
Median Technologies recently announced a research collaboration agreement with Xingtai People's Hospital, Xingtai City...
Bay Labs Completes $5.5 Million Series A Financing for AI-Driven Ultrasound
News | Cardiovascular Ultrasound | December 21, 2017
December 21, 2017 — Bay Labs, a medical technology company applying...
The Philips IQon Elite debuted at RSNA 2017 with features designed for emergency and trauma imaging. #RSNA2017, #RSNA17

The Philips IQon Elite debuted at RSNA 2017 with features designed for emergency and trauma care, as well as oncology. Image courtesy of Philips Healthcare

 

Feature | RSNA 2017 | December 21, 2017 | Greg Freiherr
The tenets of value-based medicine rippled across the exhibit floor like waves from a stone tossed in a pond.
Videos | Artificial Intelligence | December 21, 2017
ITN and DAIC Editor Dave Fornell shows several examples of how vendors are incorporating AI, deep learning algorithms
Videos | Artificial Intelligence | December 19, 2017
A post-game roundup by ITN Contributing Editor Greg Freiherr and ITN Editor Dave Fornell on the key trend of artifici
GE and NVIDIA Unveil Artificial Intelligence Upgrades to CT, Ultrasound and Analytics Solutions
Technology | Artificial Intelligence | December 14, 2017
At the 2017 Radiological Society of North America (RSNA) Annual Meeting, GE Healthcare and NVIDIA announced a series of...
Overlay Init