AI Ethics: Why It Is Important To Establish Ethics in AI Use Cases

Written by Hannah Zhou

Opening your eyes, you adapt to bright light and echoing sounds around you. Beauty enshrines you; Groups of huddled, silver-colored objects caressed your frail body. Their touches are cold but full of love. Finally, your eyes adjust, greeted by an ecosystem of machines.

“Beep, beep,” A machine buzzed, welcoming you to the world.

Yes, an ecosystem of machines. You heard me right. By the end of this blog, I will convince you that an ecosystem of machines is not such a weird idea.

Establishing Ethics in AI

Humans’ knowledge in AI is growing exponentially, at a rate that far exceeds the ability of society to catch up. Thus, a plethora of issues, especially ethical issues, arise. Today I will summarize core ethical issues surrounding business regulations, society, humanity’s future, and a key solution to decrease the ever-growing gap.

Business and Consumer AI Regulations:

It’s a race where the winner is always AI. Regulations cannot keep up with the advancement of AI, causing many problems. For years, industries have roamed freely on the Internet. They enjoy full exploitation of consumers-whether the consumer cares or not-by collecting user data to improve their algorithms and machines. In the past, parties couldn't access your private information or seek your daily ventures, but now accessing data is as simple as typing in a username. More dangerously, AI can commit crimes such as spreading misinformation, hacking systems and disclosing personal information publicly. Therefore, rapid regulations are essential.

Adding regulations most likely would prohibit AI potential. Unfortunately, adding ‘moral claims’, or adding human behavior into a machine constricts its efficiency. For example, say that regulations are established in the advertisement market. Before, AI could access your data and send advertisements favored towards the consumer. Now, AI cannot access that data, and the advertisements given to the consumer are everyday old random Ads you bore through when Cable was still a thing. Both the business and consumer lose benefits due to this regulation. Essentially, the consumer already experiences the zenith of AI’s powers, and then it’s stripped away. Essentially, a spoiled kid is given everything, and when it’s time to go to the real world, the kid is stripped of their long-lasting bliss.

According to The Harvard Gazette, the U.S is behind on data-privacy laws compared to other countries. The European Union already placed ‘rigorous’ data-privacy laws and is creating a foundational, ethical framework for AI. The U.S has not tackled this situation yet.

AI and Society:

As AI becomes more complex, its values become an important factor to evaluate. AI needs to be ingrained with values that respect community values.

An example is creating driverless cars. A driverless car cannot be implemented into our society if it does understand the laws and values of its community. What are the values? They are the simple ones, like don’t kill, don’t steal, etc, but also those that are the intricate values wrapped around the finger of law. If the AI commits a crime, who’s in charge, what punishment it gets, how can one trace the steps that led to the AI’s decision? What is the boundary between ethics and law? These questions are extremely difficult to resolve.

Another serious example is unintentional bias. If AI is fed prejudice in data and creates AI algorithms that can result in discriminating results and consequences. Similar to humans, machines learn in a certain environment, a certain ecosystem, and affect their opinions and perspective in life. A misleading environment can cause many problems. According to logically.ai, the COMPAS (Correctional Offender Management Profiling for Alternative Sanctions) algorithm created a model that predicted twice as many false positives for black offenders than as many false positives for white offenders that would become a recidivist (a person committing repeated crimes).

AI Exceeding Human Ability: (a bit more of a fun idea)

As said earlier, the tantamount growth of AI leaves a demanding question: what if artificial intelligence surpasses human intelligence? There are many interesting takes to this- perhaps possible- future.

Some assume that without proper regulation, super AI will function around rationality, rather than a mix of rationality and morality, the mind of a human. This is because some argue that rationality and morality are independent of each other. In other words, superAI will not be based on ethics, and rather factual information and efficient algorithms. It is correct to not speed to the hospital even though it’s an emergency. It is correct to save 5 people on a railroad track than to save your mother because you will save more people. It is correct to remove this entire species because it is dangerous to the world.

This idea leads to another potent suggestion: what if superAI leads to the extinction of humans? Do you ever joke around about how the world is better without humans? Corruption, global warming, global war, inequality among citizens, all these factors may lead AI to decide that the end of our existence is rational. Then, the theory in video games of machines taking over the world may just as well come true.

Although these observations and theories sound threatening, humans’ extinction is partially a philosophical approach to the future of AI. Much of it is not to be worried about. Plus, if humans can implement useful regulations and ethical components to AI, most of these possibilities will remain an old fairy tale.

A Solution to the AI Ethics Dilemma

I have a few solutions to diminish the gap between knowledge and regulation, but now I will talk about the most important answer-education. If districts implemented an AI ethics course, it will help young kids and adults understand the ethical issues surrounding AI. The younger the students, the better accustomed they will be in the future. When those students grow old and hold political power, such as the legislative branch, they can address these issues appropriately. The lag of regulations will steadily increase in pace. I believe this issue can be solved with a proper environment, a proper model, proper learning and understanding, and a proper transformation to an educated adult. Moreover, AI can teach other AI, like creating an ethics bot that trains other machines the moral ideals of human beings. An ecosystem of humans lives together with an ecosystem of machines.