	The report by Stanford University is very thought-provoking. The experts of the study panel make great efforts to carefully investigate the development and application of artificial intelligence and provide us with a blueprint imagining how our life will be in the future. However, I do not think some of the opinions stated in the report are reasonable. 

	Firstly, when the report discusses the use of artificial intelligence in public safety and security, it points out that AI could help remove the bias of humans decision-making. I totally disagree with this statement. On the contrary, I think that AI could even be more biased than human beings. For example, region A has a higher crime rate than region B. An AI decision-making system that is trained by some data including crime rate might predict that a people who lives in region A is more possible to commit a crime. If this happens, the AIs prediction would be ridiculous and unfair. Crime is such a complex problem that cannot be reliably predicted by just piling up the data. Additionally, it is hard for people to truly understand any details about how an AI model works, like neuron networks.  After training the model by a huge amount of different types of data, people can hardly control how the data of different types is combined. In this case, the bias hidden behind the result predicted by AI is much harder to be figured out. What is more, AI system gets so much data of common people. Once the data is gained by criminals, the public security would face threat. Therefore, I believe that it is irrational for people to anchor their hope on an AI program to help safeguard them. 

	Secondly, while the report holds an optimistic attitude towards the use of AI in transportation, I think that it is highly vulnerable to attack or accidents. If a slight problem occurred in AI controlling system, the behavior of vehicle might be greatly affected, like being out of control. Also, AI would be vulnerable to computer virus or internet attacks. Since AI controlling system is also a kind of computer software, it could be hacked by some ill-intentioned people, which could be detrimental to public transportation. What is more, the prevalence of self-driving car could lead to the disappearance of humans driving skills. If AI could perform daily driving tasks for people, they would not be willing to learn how to drive. As a result, they might gradually lose the ability to handle emergency cases like when the AI system does not work properly. At the same time, in order to effectively regulate all the vehicles controlled by AI, the government must build a central system to track, supervise and regulate them. Such kind of system require enormous computational resources. Based on the reasons above, I think that self-driving vehicles could not be prevalent in the near few decades. 

	In conclusion, I did not agree with some opinions in the report by Stanford University. I believe that it should still be a long time before people could fully understand the working mechanism of artificial intelligence system and thoroughly prove its reliability. Only when we are sure that AI system is reliable could we massively use AI in our daily life and vital parts of our society. 
