via Todd Richmond
The marketing promises of anything “new.” For emerging tech, turns out those broad goals are the same. Bonus points if you get all three. Another example of AI getting one of those right, in this case “faster” at generating MRIs. With the faster aspect meaning that a patient has to spend less time sitting still inside a claustrophobic noise generator.
This problem is tailor made for machine learning algorithms – there is a ton of data (existing MRI scans), there is a broad diversity of data (“normal” vs varying not normal), and the system is filling in gaps rather than creating something from scratch.
Look for AI to begin to creep into many different processes, with the first ones being similar to this – making a well understood task faster. Next will be to improve accuracy, next will be to lower costs. In the short term, not a lot of impact to human jobs. Yes, some parameters will change, but we’re not talking about replacing something a human is doing. That will come soon enough though, so part of the development should be looking at the new opportunities for “people jobs.” Particularly now given the shifts forced by the pandemic, some of which will be long term and/or permanent. As we train our algorithms, we need to be retraining our population for the next wave of workplace changes.