Challenging Stanford’s One Hundred Year Study on Artificial Intelligence (AI100)

I expect that the degree of human interaction in artificial intelligence technologies in certain domains will be a lot lesser than predicted by the study. 

For instance, in the domain of transportation, the study has mentioned that humans will become partners to self-driving cars and that there will be models and algorithms for modeling of human attention. 
I disagree with this prediction. I believe that by 2030, self-driving cars will be FULLY autonomous and there will be no element of human interaction whatsoever. Artificial intelligence in autonomous cars will have developed to such an extent that a fully autonomous car will certainly be safer than a car requiring the human passenger’s attention. Also, as the study has mentioned, the semi-autonomous approach is not sustainable as people will pay least attention to driving at the moment they are most required to do so.
The popularity of services like on-demand and shared transportation is expected to increase in the future. This ties in with the minimal human interaction component. For instance, for point to point transportation, a person wishing to go from A to B will simply book a car which will arrive at A and take the person to B. The person will not be involved in the driving whatsoever. 
Moreover, most of the benefits which the study attributes to self-driving cars, including increased comfort, decreased cognitive load, etc, are only achievable if the person travelling is not expected to pay attention to the driving.

In the domain of education, the study has stated that MOOCs and other forms of online education will become part of learning, but formal education will not disappear.
However, I believe it is possible that specific types of formal education like those for working professionals and career changers (as mentioned in the study, they do not need a high level of face-to-face interaction) will disappear. I expect that a number of courses of a specific type, (professional courses, courses which train people in specific technical skills, etc) will become fully automated, devoid of any human interaction. In other words, higher education will be completely transformed as far as professional or skill-oriented degrees are concerned.

Also, I do not completely agree with some other arguments made in the study.
In the executive summary, the study has claimed that artificial intelligence technologies are highly tailored to specific tasks.
While current artificial intelligence technologies have certainly required years of specialized research, I believe they are not domain specific. For example, there are various applications of natural language processing algorithms in different fields like bioinformatics. As far as tasks are concerned, one of the most popular artificial intelligence techniques - deep learning, is not task specific and is being applied in a wide variety of tasks, both supervised and unsupervised.

Lastly, I believe there should be some official institution which provides accurate information about recent artificial intelligence developments. For instance, a recent incident, where two Facebook chat-bots developed their own optimized language to interact, was blown out of proportion by the media. While there were articles pointing out that Facebook’s requirement was bots which could interact in human languages and so they worked on modifying the system, these articles were largely overshadowed by the ones claiming that Facebook shut down the bots fearing the rise of artificial intelligence as Hollywood’s Skynet.
The dangers of artificial intelligence should not be ignored but, as mentioned in the study, this kind of portrayal of artificial intelligence by the media will affect the public’s attitude towards beneficial artificial intelligence technologies.
