<The Prophets> is a fictional essay written by KAIST professors who are researching in each field to predict the future based on a specific point in time. It is not just a fantasy level, but because it delicately coordinates current research achievements and future feasibility, <The Prophets> will give soup subscribers a preview of the future that science predicts for as little as 10 years and as long as 50 years or more. will. (Written by Jaesik Choi, Professor at KAIST)
Intelligent systems created through artificial intelligence technology have been a staple of science fiction films and dramas in the past.
Objects that can think, think and speak on their own made with artificial intelligence technology range from robots and cars to central computers that coordinate large-scale systems.
In past sci-fi movies, there were often cases where the appearance of such an artificial intelligence system was predicted a little faster than we thought.
Here, we would like to take a look at when autonomous cars that drive automatically when a person gets in, a system that predicts human behavior in advance, and a system that automatically drives large systems with free will will come into our lives.
'Kit' becomes a reality... Self-driving car
In the late 1980s, an American drama called 'Operation Blitz Z' was very popular.
Of course, famous actors and exciting stories also played a role, but among them, 'Kit' was a car with artificial intelligence technology that allowed the main character to communicate through a watch worn on his wrist, and to judge the situation on his own.
It has a high-performance camera and an infrared night camera, so it recognizes and responds to people and situations around it, and also has the function of estimating the target person's face before plastic surgery.
In addition to multilingual interpretation, it is possible to have a natural conversation with the main character, and as time goes by, you can learn the values of the main character, such as having similar words with the main character, by understanding the tendency of the main character.
Of course, Kit jumped, flew in the sky for a short time, rode on water, rode on one wheel, or sprayed oil to defeat the bad guys.
If artificial intelligence technology develops in the next 10 years or so, some of the functions of the 'kit' above are likely to be used more in real cars.
Level 5 (complete autonomous driving), which is fully autonomous driving in all roads and driving situations, has not yet been achieved, but even now, as long as a person is responsible for driving, the level 2 (partial automation) function that automates steering and acceleration and deceleration of the vehicle is being implemented. I have many commercial vehicles.
In the next 10 years or so, it will not be difficult to achieve level 3 (automation), where cars are responsible for driving in sections under specific conditions, such as highways.
In addition, it is expected that there will be more cars driving at level 4 (advanced autonomous driving), which operate automatically without driver intervention in connection with traffic signals or parking systems so that autonomous vehicles can drive well on city roads.
In addition, with the rapid development of artificial intelligence technology that learns and processes human language, it is expected that cars will be equipped with a function to joke with people and talk about the driver's favorite values or topics while gradually learning the driver's preferences. eagerly anticipate it.
However, in the process, people believed in the functions of self-driving cars too much, and then accidents occurred in suddenly changed natural conditions such as hail and heavy rain, or the functions of autonomous cars were partially damaged due to communication failure or interference with magnetic fields, causing users to suffer greatly. I think you may experience some inconvenience.
In the process, I believe that explainable artificial intelligence technology, which can analyze how complex self-driving cars operate and provide the basis, will greatly contribute to making safe self-driving cars.
Like 'Minority Report'... Artificial Intelligence Evaluating Future Behavior
In the movie 'Minority Report', which was released in 2002, a system that predicts the future of crime will be developed 30 years from now, in 2054, giving a warning in advance before special crimes such as murder occur, so that special police It showed the future look of blocking.
Although it was not the role of artificial intelligence to predict the future in the movie, if it can predict and prevent crimes such as murder in advance, it will contribute to the safety of ordinary citizens and the creation of a safe society.
The film also showed that a good person could be turned into a criminal if the prediction system was wrong, and that a person's life could be limited by this prediction system.
Even today, artificial intelligence is being used to evaluate people's future.
It is known that US courts use the COMPAS system with artificial intelligence technology to predict the prisoner's probability of recidivism when deciding on a prisoner's parole.
In other words, an artificial intelligence system predicts whether an inmate will commit another crime in the future, and affects whether or not he is released on parole.
So, what if a system like this isn't perfect?
In fact, it has been revealed that the compass system makes biased predictions that differ by race, and it has become a big social issue that needs to be corrected.
This can be seen as a case in which it is important to analyze internal decision-making and ensure that there is no unfairness from the human point of view in the case of artificial intelligence systems that can predict human behavior and restrict their activities.
Amazon in the United States is known to use an artificial intelligence system during the recruitment process to compare the resumes of new employees with those of existing employees who are performing well to select employees who are more suitable for the company.
Japan's Septeni Holdings reported that by quantifying the performance of each employee, the amount of data per employee was about 180 for new employees and 800 to 1,000 for 10 years at the company.
By using this big data for learning of the artificial intelligence personnel evaluation system, it was reported that the performance of employees can be specified and referred to in personnel evaluation, and the rate of resignation can be predicted and the rate of resignation can be reduced by assigning them to more suitable tasks or departments in advance. .
As a result, it looks positive if you select employees who are well suited to the company and predict the employees who will leave the company in advance to arrange departments.
However, if the company I work for uses such a personnel evaluation system, and if the result is something that is difficult to accept as it is (e.g. I have never thought of that, I predicted that the employee will leave soon), how can I resolve this resentment? ?
However, within the next 20 years, there is a possibility that many companies will digitize the work performance of their employees and analyze this big data system with an artificial intelligence system to refer to as auxiliary data for personnel evaluation.
Even now, in leading overseas professional soccer leagues, the online artificial intelligence system analyzes each player's activity and record in real time and predicts the condition and result of the players even before the game, so it is only a matter of time before this applies to ordinary office workers. I think it might be.
However, even in this case, the artificial intelligence system may not be able to consider the personal situation that can be considered in the existing HR system, in which a person decides whether or not the AI system is wrong.
So here, too, it is important to be able to transparently analyze and disclose the decision-making of artificial intelligence systems.
Realization of 'HAL 9000'? Artificial intelligence that coordinates large-scale systems and opposes humans
If you look at the movie 'Space Odyssey 2001' directed by Stanley Kubrick released in 1968, there is a computer called 'HAL 9000' that thinks for itself, controls the system, and talks to people.
In the movie, it says it went into operation at the University of Illinois at Urbana-Champaign in 1992.
I entered the Department of Computer Science at the University of Illinois in 2005 and graduated with a PhD in 2012, but unfortunately I have never seen such a computer.
Instead, in 2006, I was able to see that an alumni of the University of Illinois of the same age sold Youtube to Google at a value of about 2 trillion won.
Of course, the value of Youtube is thought to be several dozen times greater than that.
**If the 'Go to View' button is not pressed, you can view the address by moving it to the address bar.