	Humans interest of artificial intelligence is growing rapidly these days. More and more resources and efforts are put on the study and development of artificial intelligence. According to the Stanford 100 years report, although our study on artificial intelligence is still at the beginning stage, we have already proposed some domains where technologies generated from artificial intelligence can be used at. These domains include transportation, home/service robots, healthcare, education, low-resource communities, public safety and security, employment and workplace and entertainment. The report provides a lot of facts of the current status of some artificial intelligence technologies and also expectations of their future development. It is definitely true that we have made a great start on artificial intelligence but some of the expectations are too optimistic. There are some gaps between artificial intelligence and a real persons brain and these differences may never be eliminated. 
	According to the report (p.20), the self-driving car technology may lead to a complete change in transportation in the future. For example, the performance of self-driving cars will exceed the performance of human drivers in the future and human drivers will be replaced completely. As a result, traffic accidents will decrease and people may never need to buy their own cars when artificial intelligence algorithm can arrange everyones traffic demand efficiently with public transportation. However, this future is very difficult to reach and it may never come true. First of all, for some people, driving is not a burden or a necessary tool of traffic but a personal interest and hobby. These people will probably drive by themselves even when their cars have self-driving technology. In this circumstance, these vehicles are more difficult for AI drivers to calculate. There comes the most important concern of AI drivers: are self-driving cars truly safe. The development and preparation of self-driving cars have started since 2005 and Googles and Teslas self-driving cars have already traveled in cities. The self-driving AI seems to be very accurate and reliable with both the complex algorithm to analyze the traffic situation behind the AI and the information gathered through sensors placed in the cities (p.22). However, there are many possible situations which may happen on the road and AI cannot respond to these accidental events as reasonably as human brain does if these events are not anticipated by programmers. For example, in 2016 May 7, a Tesla Model S self-driving car crashed on a white truck during the Autopilot mode and the driver on Tesla is dead. The possible explanation to this tragedy is that the camera of autopilot system cannot gather the full view of the truck since it is large while the truck is white, therefore the autopilot system considered the truck as cloud or just simply ignored it. The real problem lies beneath this accident is that the database of self-driving AI to identify objects is too small compared with all the possible situations in the real world. The influence of this problem may be reduced if we keep improving self-driving AI and its database. However, the potential danger will always exist.
	Other than the safety concern, another potential problem of self-driving cars is private information. If the self-driving system wants to reach the greatest efficiency as the report suggests, it is necessary for the whole network to gather both traffic information and peoples transportation plans. There comes possibilities that those personal information, including traffic plans and home addresses, may be released since they are all accessed by self-driving AI.
