Artificial Intelligence Ethical Issues

When we discuss artificial intelligence, it is almost impossible to avoid talking about the controversies and ethical issues that come with the topic. Artificial Intelligence (AI) has helped the world so much today and there are even more innovations and use cases in this space. However, apart from the immense benefits of AI, a lot of artificial intelligence ethical issues stare us in the face daily. Discussions about artificial intelligence ethics are on the front burner of many individuals, researchers, and organisations.

Let us consider the major areas of concern in using artificial intelligence.

Major artificial intelligence ethical issues.

1. Artificial Intelligence Bias

Bias is the major ethical issue of artificial intelligence systems or machines. We train intelligent machines based entirely on the datasets and assumptions of the developers or engineers. We can hardly eliminate human bias, no matter how much we try. Arguably, the data and configuration of data used to train any machine may depend on what the creator expects the machine to do after training. That means, for instance, that a machine can learn to discriminate between people if it finds such a pattern in the dataset presented to it during training. Often, user-generated data is more biased than data generated by machines or industrial field instruments. The question now is: Is it possible to eliminate human bias in machine learning when most of the data we currently have today is full of bias?

2. Artificial General Intelligence (AGI)

As opposed to building intelligent machines to handle specific tasks, some organizations are trying to build AI agents that can handle all tasks that a human can with ease. We expect such agents to behave exactly like us and even exceed human intellectual abilities based on the wealth of data they possess, as well as outstanding processing speed. When we build systems that are more intelligent than we are, shall we co-exist peacefully? The implications of having these kinds of machines are far-reaching. (You can find a deeper, more practical analysis of AGI here)

3. Deep Fakes

Society is slowly but surely losing trust in the media because of this area of artificial intelligence. Deep fake entails using images and voice to mimic media characters to as much precision as possible. This makes it extremely difficult to distinguish between a real human newscaster, for example, and a fake one. The ripple effect is that no one trusts the media anymore, as we can hardly believe each other. After all, how sure am i that I am interacting with an actual human? Not that we cannot have good use cases for deep fakes, the challenge here is that they are getting proliferated for personal and political reasons. With deep fakes, the future of our truth hangs on a wobbling balance.

4. AI-assisted Human Judgement

Data analytics has become a great tool for any sector of the economy. Most organisations use it to make informed decisions. This is an area of artificial intelligence and the implication is that most human decisions are now guided by the insights gotten from analytics. This can pose a threat when the process of analytics is abused for some sort of personal political gains by unscrupulous entities. 

5. AI-Powered Machines of War

Many developed countries are in a race with each other to stand out both in weaponry and in response to potential wars. They have applied various fields of artificial intelligence to build complex war machines. The issue here is that most of these weapons are designed to act autonomously. Do the machines know exactly when to stop? In identifying targets and decision making, are these machines able to act for the common good of man? When we put human lives entirely into the hands of machines, we do not know what these machines may come up with in an event that they meet a strange scenario. Humans may use common sense to analyze situations, the logic that machines have can become messed up in complex situations that they have learnt nothing about. We can be sure that no one accurately predicts the flow of events in a scene of war.

6. Privacy Issues

Surveillance systems capture more information about us than we can imagine. While most websites request you accept cookies, they have those cookies that they feel are compulsory. Most times, users are not aware of what and how much information about themselves they are giving out and how this data is used. Some tech giants and platforms have more information about us than we even have about ourselves. And then, they use the data for analytics and apply their findings to various areas. Today, personalised ads are becoming frustrating for users because the ‘tools’ that collect and analyze data for recommendation systems are becoming smarter and taking in the minutest details about us for their ‘jobs’. This gives us the feeing that we’re being monitored.

Conclusion

As we build and deploy more artificial intelligence systems and machines today, we must understand that we are allowing these machines to take important decisions. That said, it really matters what these machines know and think. Their impacts on the society also matters a lot. We are responsible and accountable for whatever these intelligent systems do. So, we should act responsibly and be transparent in the collection and use of data to train machines. Transparency is also highly important in using algorithms for machine learning. There are also questions about the right of robots in the future (Anything is possible in this world. Let’s not underestimate human idiosyncrasies). Ultimately, when we build artificial intelligence systems, we should ensure that we are creating machines, systems and agents that will facilitate and preserve the existence, sanity, and overall well-being of humanity.