The previous article mentioned the current application of artificial intelligence (AI), particularly in the field of machine learning. Benedict Evans summarizes this technology as a series of artifacts that, after gathering enormous amounts of information, can perform seemingly impossible tasks like image recognition or autonomous vehicles. A good example is the following comic from 2014, just before Deep Learning took off :
These Machine Learning applications will continue to increase in quantity and complexity, as more and more professionals, entrepreneurs, and investors will devote their attention to these fields. The Organization for Economic Cooperation and Development (OECD) published a study in which it mentions that from 2011 to mid-2018 around 50 billion dollars of private capital had been invested in AI, and the trend is on the rise.
An example to understand the evolution of these technologies is Watson, a software designed by IBM to answer questions in natural language. In 2011 Watson managed to defeat Jeopardy ! two of the best players in history and is currently capable of performing medical diagnoses.
Kevin Kelly, in his book The Inevitable: Understanding the 12 Technological Forces That Will Shape Our Future ( 2017), tells an anecdote in which Watson was able to correctly guess a type of intestinal infection that he had suffered from the description of the symptoms. Kelly quotes Alan Greene, Scandau Medical Director, who comments that “at the rate AI is improving, a child born today will rarely have to see a doctor during adult life for a first diagnosis.”
Although not all of us have access to Watson, IBM already counts several pharmaceutical companies among its clients. Many of these, using Watson’s medical intelligence, provide personalized advice to consumers with chronic conditions based on their prior treatment history. It is not difficult to imagine other industries where similar innovations could be developed. Kelly mentions some examples:
Biomedical sensors that collect data are used to generate personalized treatments and even continuously adjusted.
Bringing sellers and buyers together using AI, similar to what social networks or online stores do. From this, generate personalized financial packages to the needs of each client.
Create music in real-time for different scenarios (events, video games) which change according to the specific circumstances of those scenarios.
Software that allows considering external factors such as weather forecast, airport delays, accidents, and changes in the value of currencies.
Advertising campaigns that can detect the consumption pattern of users in such a way that they seek to attract the attention of those who spend the most.
Reviewing this list we can see that primitive versions of these innovations already exist. For example, the Apple Watch already has certain technology to capture biomedical information, or social networks can collect a large number of data points from a user. These innovations are hard to see as primitive applications, but they will seem that way for decades to come.
As we saw in the previous article, the Machine Learning process requires a large amount of data and wearables like the Fitbit or the Apple Watch as well as the devices that will be part of the Internet of Things (smart appliances, monitoring technologies) will contribute. for the data sample to continue to grow, making these algorithms more powerful.
It is difficult to imagine any industry that is not going to be affected by Artificial Intelligence. The clearest example is drivers and the manufacturing industry, due to the development of autonomous vehicles and robots that will be able to move safely among their human “companions”. Today the robots used in the industry tend to work separately from people, which limits their productivity.
The robots of the future, in addition to the strength, speed, and precision with which they already have, will be able to interpret the images and sounds of their environment.
Interacting safely with people, and will also be connected to a network (either the Internet or a local network) from which they can obtain specialized information to carry out their tasks.
White-collar jobs will not be safe either. Many modern jobs consist of detecting, interpreting, consolidating, and translating information from one context to another. We are part of a value chain through which information flows. The development of Machine Learning, specifically Deep Learning, has exponentially expanded the capabilities of machines to process information.
Our current tools ( gadgets and software ) will gradually become more powerful. These tools will perform a greater number of tasks, each time more complex, and with less human supervision. This is industries as diverse as Restaurants, Agriculture, Electronic Commerce, or Financial Services.
Kevin Kelly states that before the end of the 21st century, 70% of today’s occupations will be able to be automated, reflecting what happened during the Industrial Revolution. Likewise, new jobs are expected to emerge, although the skills that those jobs will require will be different from the jobs that will disappear.
In the words of Kai Fu Lee, founder of Sinovation Ventures and former CEO of Google China: “Artificial Intelligence can optimize, but not create,” so the most creative jobs are the safest. That is an activity that is beyond the synthesis capacity of machines. Some characteristics that define these activities are:
Warmth, Interpersonal Attention
As customer service becomes automated in the service sector, many could pay a premium for interpersonal care.
Creativity and strategy
By definition, a computer cannot think outside the box. That is, even though there are computers that apparently can learn anything (like Watson), they move within the parameters in which they have been programmed. Just as the term Deep Learning describes well the capabilities of machines, for people we could use the term Broad Learning.
Design of personalized experiences
Although technology allows the personalization of products, diets, etc., there will still be many niches available in tourism, event planning, or any other activity where human sensitivity cannot be replicated by machines.
Kai Fu Lee shares his optimistic view of the impact of Artificial Intelligence in his Ted Talk “ How AI can save our humanity”. Lee proposes that as machines embrace more routine tasks, people can focus on more humane tasks, occupations where creativity, empathy, and compassion are essential parts of the job.
Lee cites a PwC study that estimates a $ 16 trillion increase in Gross Domestic Product (GDP) around the world due to AI. Faced with this possible scenario of high economic growth combined with the loss of many jobs, experts such as Martin Ford have proposed considering Universal Basic Income as a viable alternative, although they recognize that it is not a panacea and that its implementation would have to be carried out with the proper incentives.
This proposal is already part of the discussions of the candidates for the Presidency of the United States for the 2020 elections.
In summary, Artificial Intelligence is here to stay and will gradually spread throughout all industries, attracting all those tasks where its capacity for synthesis and learning makes supervision or human intervention unnecessary.
But still leaving room for those tasks that are essentially human, and also leaving a huge challenge to society in the face of possible economic inequality that could affect us in the following decades.
In the next and final article, we will talk about the possibility of advances in artificial intelligence beyond machine learning and the possibility of going beyond human intelligence.