AI Implementation and Future Directions in Canada

[brolias]

Hi there, just wondering what the current implementation for Radiology AI is across the major centers in Canada are currently? Also, future directions which may be different than America? 

Wondering if eventually will reduce radiologist demand...

[1D7]

This is a somewhat complex topic but the attitude of many computer scientists and some physicians outside of radiology is that it is mostly a pattern recognition field and thus will be disrupted or replaced by 'machines'.

The attitudes of most radiologists boil down to extrapolation from historical trends in AI development, previous technological increases in efficiency, and the current state of the field. Mostly this has been positive for radiologists, or at least has resulted in increased demand for radiologists. AI is not a very interesting topic for radiologists as the area of AI development is filled with questionable data/methodology. It is also a tired topic that yields little fruitful discussion, as most computer scientists and non-radiology physicians don't have a very good understanding of our role and our impact on patient care. Thus far the implementation of AI is either limited or low impact. Breast CAD is not useful though widely available; RapidAI saves just a few minutes for the stroke team but still requires detailed review by a physician; AI head bleed detectors are wrecked by artifact; and ML predictive text populators are still in their infancy.

Most radiologists are hopeful that better AI tools will be developed soon to increase our efficiency, as we are relied upon more & more every year for diagnostics & procedures.

[Asphalts]

From what I have heard from my vampire colleagues, they are not too worried about AI decreasing job prospects for many years to come. In fact, it will likely only enable radiologists to do their job more efficiently and better!

But of course eventually everything might be replaced by AI who knows