The right and wrong way to use artificial intelligence – New York Daily News

For decades, scientists have been giddy and citizens have been fearful of the power of computers. In 1965 Herbert Simon, a Nobel laureate in economics and also a winner of the Turing Award (considered The Nobel Prize of computing), predicted that machines will be capable, within 20 years, of doing any work a man can do. His misplaced faith in computers is hardly unique. Sixty-seven years later, we are still waiting for computers to become our slaves and masters.

Businesses have spent hundreds of billions of dollars on AI moonshots that have crashed and burned. IBMs Dr. Watson was supposed to revolutionize health care and eradicate cancer. Eight years later, after burning through $15 billion with no demonstrable successes, IBM fired Dr. Watson.

In 2016 Turing Award Winner Geoffrey Hinton advised that We should stop training radiologists now. Its just completely obvious that within five years, deep learning is going to do better than radiologists. Six years later, the number of radiologists has gone up, not down. Researchers have spent billions of dollars working on thousands of radiology image-recognition algorithms that are not as good as human radiologists.

(Iaroshenko Maryna/Shutterstock)

What about those self-driving vehicles, promised by many including Elon Musk in his 2016 boast that I really consider autonomous driving a solved problem. I think we are probably less than two years away. Six years later, the most advanced self-driving vehicles are arguably Waymos in San Francisco, which only operate between 10 p.m. and 6 a.m. on the least crowded roads and still have accidents and cause traffic tie-ups. They are a long way from successfully operating in downtown traffic during the middle of the day at a required 99.9999% level of proficiency.

The list goes on. Zillows house-flipping misadventure lost billions of dollars trying to revolutionize home-buying before they shuttered it. Carvanas car-flipping gambit still loses billions.

We have argued for years that we should be developing AI that makes people more productive instead of trying to replace people. Computers have wondrous memories, make calculations that are lightning-fast and error-free, and are tireless, but humans have the real-world experience, common sense, wisdom and critical thinking skills that computers lack. Together, they can do more than either could do on their own.

Weekdays

Catch up on the days top five stories every weekday afternoon.

Effective augmentation appears to be finally happening with medical images. A large-scale study just published in Lancet Digital Health is the first to directly compare AI cancer screening when used alone or to assist humans. The software comes from a German startup, Vara, whose AI is already used in more than 25% of Germanys breast cancer screening centers.

Researchers from Vara, Essen University and the Memorial Sloan Kettering Cancer Center trained the algorithm on more than 367,000 mammograms, and then tested it on 82,851 mammograms that had been held back for that purpose.

In the first strategy, the algorithm was used alone to analyze the 82,851 mammograms. In the second strategy, the algorithm separated the mammograms into three groups: clearly cancer, clearly no cancer, and uncertain. The uncertain mammograms were then sent to board-certified radiologists who were given no information about the AI diagnosis.

Doctors and AI working together turned out to be better than either working alone. The AI pre-screening reduced the number of images the doctors examined by 37% while lowering the false-positive and false-negative rates by about a third compared to AI alone and by 14%-20% compared to doctors alone. Less work and better results!

As machine learning improves, the AI analysis of X-rays will no doubt become more efficient and accurate. There will come a time when AI can be trusted to work alone. However, that time is likely to be decades in the future and attempts to jump directly to that point are dangerous.

We are optimistic that the productivity of many workers can be improved by similar augmentation strategies not to mention the fact that many of the tasks that computers excel at are dreadful drudgery; e.g., legal research, inventory control and statistical calculations. But far too many attempts to replace humans entirely have not only been an enormous waste of resources but have also undermined the credibility of AI research. The last thing we need is another AI winter where funding dries up, resources are diverted and the tremendous potential of these technologies are put on hold. We are optimistic that the accumulating failures of moonshots and successes of augmentation strategies will change the way that we think about AI.

Funk is an independent technology consultant who previously taught at National University of Singapore, Hitotsubashi and Kobe Universities in Japan, and Penn State, where he taught courses on the economics of new technologies. Smith is the author of The AI Delusion and co-author (with Jay Cordes) of The 9 Pitfalls of Data Science and The Phantom Pattern Problem.

Read the original post:

The right and wrong way to use artificial intelligence - New York Daily News