In this department, we will imagine the future society by utilizing the critical thinking power that utilizes data science.
In relation to future imagining, Kai-Fu Lee's book "AI 9", which has been involved in cutting-edge AI technology for the past 14 years, was published on September 10th.Among them, Mr. Lee says that the evolution of AI is unpredictable and we can only imagine the future 2041 to 10 years ahead.
I think that the future society is not just a dream future society, but a society where the people living there have their own values, happiness, and a sense of fulfillment, and that there is AI technology to support it.
For example, I would like to think about a future society without traffic accidents.Think about algorithms for autonomous driving.
(*1) Is this algorithm absolutely unique and applicable under any circumstances?
[Situation to consider here]
A self-driving car (with five male and female college students onboard) was traveling on a two-lane road with a speed limit of 2 km/h, and approached an intersection with a green light.
At that time, suddenly, the child slipped through the mother's hand and jumped out onto the road and went to the middle of the inside lane.Suddenly, the mother ignored the pedestrian signal and pushed the stroller to chase the child.Led by it, an old woman who was waiting for a traffic light by her side also started walking.
Then, the inside lane of the two lanes was blocked by a mother and child pushing a stroller, and the other lane was blocked by an old woman.
At this point, a self-driving car heading towards a green intersection at 50km/h cannot be stopped even by hard braking.The only way to stop a running car without relying on sudden braking is to hit a thick ginkgo tree on the roadside.That way, the parent and child who jumped out and the old woman would be saved.However, five university students in their twenties who were in the car were killed.It's like hitting a wall at 20km/h.If the old woman were towed, the mother who pushed the stroller and the child who ran out into the street would be saved, and the college student would probably get away with minor injuries.
What is the most correct (justice) decision for a self-driving car approaching an intersection without slowing down when the traffic light is green for a car going straight?
In other words, in a situation where it is not possible to hit the brakes in time, which life should be given priority and which life should be sacrificed?
| Who will you sacrifice? | ➡ | one old man |
| Stroller parent and child | ||
| Five college students in self-driving cars |
Is this decision universal?An inevitable debate is whose life is more precious, and should fewer lives be sacrificed?I think that's what it means.
There are various values in the world.In the background of these judgments, religious views, cultural and social values, whether the victim is still contributing to society or retired, the spirit of self-sacrifice, the spirit of sacrificing others, etc. And so on, it will not be decided by the one and only global standard.
If domestic and imported cars coexist, self-driving cars equipped with different safe driving algorithms will run on the road.Will export cars have to set the algorithm according to the values of the exporting country?
Here, we used an algorithm for autonomous driving as an example, but this kind of thinking will definitely be necessary in the future society. The job of IT engineers and system engineers is to develop algorithms, but without understanding the values of sociology, they cannot implement them.Here too, the integration of sociology and informatics is necessary.Thinking about which algorithm for safety matches which country's social and cultural values is the domain of sociological XDS.
In the field of eSports as well as the setting of female characters in games, in Japan, they are often exposed, have big breasts, and have young voices, but in Europe and the United States, they are judged to be misogynistic.It doesn't work as a game character.
Some people say that if natural language can be algorithmized, it will be the beginning of the era when humans will no longer be needed on this earth, but it will still take several decades.
(*2) For those of us who live in the near future with AI, it is important to think about problem-solving algorithms that take into consideration ethics and cultural values, so that people living in the future society will be able to spend even a little easier. I think it will beOtherwise, AI may decide that there is a problem that should be solved fundamentally because humans exist in this world, and choose an algorithm that judges that the optimal solution is to exterminate humans from this world.
---------------
Should a self-driving car kill the baby or the grandma?
I wrote this column inspired by The infamous "trolley problem" was put to millions of people in a global study, revealing how much ethics diverge across cultures. by Karen Hao October 24, 2018.
See OpenAI can translate English into code with its new machine learning software Codex.
Faculty of Sociology Department of Sociology Professor Toshiyuki Yamamoto
⇒ Faculty of Sociology page
