Ethics of Artificial Intelligence
In an interesting incident at Facebook, the artificial intelligence systems at Facebook developed a language to communicate with each other that could not be understood by the very engineers that made the system. Facebook immediately shut down the system.
Elon Musk and other top scientists and entrepreneurs have expressed concern over the need for rules and regulations for AI before it gets too late. Elon Musk has suggested three regulations to the government.
1. Firstly, the same laws should apply to an AI system that apply to its operator. In other words, an owner/operator of an artificially intelligent system should not be able to get away for a crime that his machine committed.
2. Secondly, an AI system must clearly disclose that it is not human.
3. Lastly, an AI system should not be able to retain or disclose confidential information without the explicit permission from the source of information.
With the revolution of artificial intelligence, it is a risk that the synthesized intelligence might not always be ethical and robots with this unethical intelligence might become much more powerful and harm their own creators. While, it is important to control AI systems with rules and regulations, ethics of artificial intelligence is also a topic of concern. We need to know what ethics are to incorporated into systems that are artificially intelligent and what ethics do we need to have while developing these systems.
One important aspect of the ethics of artificially intelligent systems is privacy. If AI systems learn to interpret our language they can listen/read our conversations and report them to a perpetrator which might again be a human or another artificially intelligent system. With the development in technology where data is collected from our phones on a minute to minute basis, we cannot afford for AI systems breaching our privacy. Privacy is an integral part of living a meaningful life and hence needs to be strongly incorporated into the ethics of AI.
Also, Joseph Weizenbaum was a computer scientist who argued in 1976 that AI should not be given positions that require human understanding, respect and care such as a judge or a police officer. Replacement of such positions with machines will put us in a very alienating and frustrating position. While, some people have argued that an AI system might be more impartial and just, it is difficult to rely on a machine for some jobs. Human understanding is accompanied with care and emotions that cannot be expected from an AI system.
Interestingly, some ethical paradoxes are also concerning builders of artificially intelligent systems because these ethical paradoxes have heavy consequences. One of these ethical paradoxes is that if an innocent person needs to be killed to save the life of five innocent persons, is it ethical to deliberately kill the innocent person. While, this paradox has been discussed for a long time, one does not expect to face such a situation in his life. With the development of AI, we need to take a stand on these questions. This paradox is very important in the development of self driving cars as in a complex situation a self driving car might face such a situation.
Ethics of artificial intelligence will shape the world in the coming years as artificially intelligent systems start to dominate the world. On a pessimistic note, we also stand at a risk of AI learning unethical intelligence from humans itself. Ethics of these systems will eventually decide if we end up having a war with something that we created or choose to live in harmony with them.
Article by Ishan Goel and Sukant Khurana
