Opinion | Why it is unlikely that AI would be a threat to humans4 min read . Updated: 13 Jun 2019, 12:03 AM IST
Early treatment of diseases may improve our lifespans as AI evolves and finds more use in medicine
The other day, we came across a quote by SpaceX founder Elon Musk, stating that “Artificial Intelligence is a fundamental risk to the existence of human civilisation". While a lot of concerns like this have been voiced regarding AI becoming dangerous to humans in the future, and while we don’t know yet whether that will be true or not, from the way things are proceeding currently, it can be said that it is least likely that AI will probably destroy humans.
Artificial Intelligence (AI) is playing a huge role in the detection and, in turn, finding of antidotes for many of the maladies that are affecting almost 95% of the world population. Through a recently developed deep learning (DL) algorithm—Convolution Neural Network (CNN)—diseases and ailments can now be effectively detected and identified.
Similarly, the International Committee of the Red Cross has developed dashboards using AI to capture large volumes of data to support its wide-ranging humanitarian work. For example, there is significant interest in technologies that could improve identification of missing persons in an effort to reunite family members separated by conflict. It is also exploring the use of AI to map population density in support of infrastructure-assistance projects in urban areas.
Medical practitioners are now able to spot various diseases like Parkinson’s, breast cancer, hole in a human organ, etc., in a more effective way. Initially, radiologists and other practitioners were manually scanning medical images for the presence of any ailment. The identification was based on their knowledge and skills. Although accurate, if, by using sophisticated machine learning (ML) algorithm, greater precision could be achieved, there are chances that more cases would be detected earlier. Recent studies have shown that DL techniques can read untoward signs of the body better than humans. It can plug the loopholes that humans cannot. Also, using technology is faster and eliminates the chance of errors in areas where humans may be prone to them.
Although this does not mean that implementing the DL algorithm will rid the world of diseases completely and immediately, instances will be reduced, detection will be faster and categorization will be better. Implementing such a technology can prove to be a breakthrough in the field of medical diagnosis. As a consequence, better disease detection and timely treatments can bring down mortality rates and increase life expectancy rates with time. DL algorithms together with nanotechnology, drug development and other genetic manipulation techniques will increase our lifespans. As American author and futurist Ray Kurzweil says, “We’ll get to a point about 15 years from now where we’re adding more than a year every year to our life expectancy."
In the Indian context, there has been a significant increase in life expectancy over the years. Interestingly, of several factors impacting life expectancy, independent research shows that the most statistically significant indicator of life expectancy is access to sanitation. A one-year increase in life expectancy is associated with every 0.25% of the population with access to modern sanitation facilities. Not surprisingly, there’s a strong negative correlation between sanitation and mortality (India’s infant mortality rate fell to 33 in 2017 from 42 in 2012). Poor sanitation conditions increase the risk of sharing life-threatening contagions. Hence the idea of building toilets in a country like India has indeed significant long-term implications. And once sanitation improves, we can concentrate on healthcare (read Ayushman), education, adolescent fertility and so forth.
In fact, by 2045, humanity will have an option to choose whether to die or not. Or at least genetic engineers José Luis Cordeiro and Cambridge mathematician David Wood so believe, as stated by them during the presentation of their book The Death Of Death. They say in their book that death will be optional and it will be possible in the next 27 years—by 2045.
The collaboration of man and machine can lead to even greater achievements. Instead of working in silos, if AI and humans join hands, the disease detection accuracy will improve. A hybrid workforce with inputs from both human beings and technology will give superior results than what either group would have been able to achieve individually.
An example can be AI-driven recommendation technologies that analyse preferences of users to predict and present the content, products and other results that the user is more likely to drift towards (think of Google searches and Facebook content). This example shows how humans providing inputs at the initial stage can help machines build their ability to analyse and act on information more independently as time passes.
However, this does not mean machines will stop needing humans. We believe the future will witness the emergence of a more symbiotic form of multiplicity. There would be asymmetry in the labour market, with simultaneous demand for high-skilled manpower (those who create the machine) and low-skilled labour (those who run the machines). Will then the demand for middle-skilled (semi-skilled) labour become minimal? Only time will tell, but the need for skilling increases manifold.
Soumya Kanti Ghosh & Prithwis K. De are, respectively, group chief economic advisor, State Bank of India, and a London-based Artificial Intelligence researcher
These are the authors’ personal views