IIT_CS480_Assignments_Week12_Participation

IIT CS 480 - Introduction to Artificial Intelligence
Assignments_Week10_Participation
github: https://github.com/IIT-MAI-2020Fall

---

This week’s discussion is about AI safety, ethics, robustness, and potential harms and benefits of AI. Here is a blog post and a paper that I’d like you to read.

本周的讨论是关于人工智能的安全性、伦理、健壮性以及人工智能的潜在危害和益处。这里有一篇博客文章和一篇论文，我希望你们读一读。

1. https://jsteinhardt.wordpress.com/2015/06/24/long-term-and-short-term-challenges-to-ensuring-the-safety-of-ai-systems/
2. https://futureoflife.org/data/documents/research_priorities.pdf

Two questions

1. Share your general thoughts on this topic. This is an open question. You can focus on any aspect of it you like. I need you to write at least one paragraph (roughly 200 words or more) on this question. This should be a general discussion of this topic; do not focus on a single problem. Next question asks you to focus on a single problem.
2. Pick one problem that you think is super important (e.g., automation and unemployment, privacy, autonomous weapons, self-driving cars and accidents, fairness, …) It does not have to be from the assigned readings. You might think all problems are super important; I need you to focus on one for this question. Share your thoughts on that problem. This is an open question. You can focus on any aspect of it. I need you to write at least one paragraph (roughly 200 words or more) on this question.

两个问题

1. 分享你对这个话题的总体看法。这是一个悬而未决的问题。你可以专注于你喜欢的任何方面。我需要你写至少一段(大约200字或更多)关于这个问题。这应该是对这个主题的一般性讨论;不要专注于单一的问题。下一个问题要求你专注于一个问题。

2. 选择一个你认为非常重要的问题(例如，自动化和失业，隐私，自主武器，自动驾驶汽车和事故，公平，……)它不一定来自指定的阅读资料。你可能认为所有问题都非常重要;这个问题需要你专注于其中一个。分享你对这个问题的看法。这是一个悬而未决的问题。你可以专注于它的任何方面。我需要你写至少一段(大约200字或更多)关于这个问题。

---

Q1：
The rapid development of artificial intelligence is bound to introduce new problems.

1. Many employees in the service sector worry about job losses – robotic arms replacing assembly line workers, agricultural drones replacing farmers spraying pesticides, robots replacing waiters. But I think there is no need to worry about this, just like the automobile was born, the carriage was eliminated, but the coachman will change his profession, and a new post will be created – the car driver. Ai has killed off some jobs in services, such as telephone customer service, but has brought in a new one, data markers.

2. Faulty optimization is also a concern. I’ve heard this description before. If you put a self-learning robot to make nails, its purpose is to make nails, and in the end it may even drain the world’s natural resources just to make nails. Although it is excellent as a nail-making robot, it has sacrificed the world’s natural resources. I think it’s a bit fanciful. This problem can be avoided by adding appropriate performance evaluation conditions.

3. About opaque systems. I do think this is a serious problem. Right now, we’re surrounded by streams of information, receiving endless messages that algorithms are pushing at us. These algorithms may seem designed for us based on past user behavior, but we’re in the control of algorithms.

---

Q2：
I want to share the ideas of system stability.
The more complex a thing is, the less stable it is.
The more advanced a thing is, the less likely it is to survive.

1. From a biological point of view, in the process of biological evolution, the more advanced the biological evolution is, the weaker its ability to resist risks to the environment will be. Microorganisms with simple cellular structures, such as the water bear worm, can even survive in space for a while. But mammals, a change in temperature, a loss of water, can kill them.

2. From a military perspective, in the Stone Age, if human society had gone to war, hundreds of people would have died. But when it comes to nuclear energy, a nuclear weapon can easily destroy a country if it makes the slightest mistake.

3. From the perspective of software design, generally speaking, the more people involved in a software project, the more complex the code, and the more bugs there are in the system. This is why agile development is a rapid iteration of its own project management model.

4. From the perspective of an enterprise, the more redundant its organizational structure, the more likely it is to fail to follow the trend of The Times and be eliminated. Must be keen to grasp the market direction, timely adjustment.

5. From the perspective of artificial intelligence technology. The more advanced a technology becomes, the more likely it is that it will only be applicable to certain segments.

To solve the problem of stability, the only thing we can do is to keep the system iterating quickly.