Exploring the ethics of AI

By

The ethics of Artificial Intelligence or in short ethics of AI is a topic that is often debated and it’s one that will no doubt continue to be discussed as artificial intelligence (AI) becomes more prevalent in our lives.

With the ethical questions and the ethics debate come many questions:

  1. AI ethics: Why are people paying so much attention to it?
  2. How can we maintain fairness and avoid unintended bias?
  3. How do we allay fears about loss of control or redundancy?
  4. What are the three major areas of ethical concerns?
  5. Considerations for now and then

This article offers insight into these ethical questions by exploring ethical aspects, ethical dilemmas and ethical implications in general, as well as the ethics principles surrounding AI specifically, and some strategies for enforcing compliance.

We all want to do the right thing. Ethics is a system of moral principles that helps us live up to our own ideals and values, as well as those in society.

The ethics of Artificial Intelligence refers not only to how we design and program these algorithms (or what’s commonly called ‘machine ethics‘), but also covers economic implications, data ethics, safety and decision making.

 

A robot with skis and machinery

 

AI ethics: Why are people paying so much attention to it?

The ethics of AI has been a topic for debate since the creation of robots and now with the meteoric rise in popularity of AI algorithms we are all wondering how to achieve an ethical outcome from these technologies as well as how to prevent unintended negative consequences that may arise due to their use.

AI is a computer program. It is a set of instructions and data that tell the machine what to do. This can be as complicated or simple as we want it to be in order to achieve a desired outcome, for example playing chess. AI has no ethics built into it so will play according to its programming without consideration for fairness or ethics, whereas an AI algorithm will consider ethics when considering things like fairness, bias or redundancy.

The goals of artificial intelligence include learning, reasoning, and perception. Today, AI is used across different industries including finance, healthcare, pharmaceutical and retail among others. Weak AI tends to be simple and single-task oriented, while strong AI systems carry on tasks that are more complex and like human beings.

The ethics of Artificial intelligence, as a topic, get a lot of attention and have become increasingly important as AI becomes more prevalent in our lives, for example with the rise of self-driving cars and other autonomous technologies that we are now using on a daily basis. With this new level of responsibility comes an increased need to ensure fairness and avoid unintended bias.

Tesla car driving at high speed

How can we maintain fairness and avoid unintended biases?

Artificial intelligence (AI) provides unique tools which enable us to do things that are impossible in a world without it. It can provide new insights, predict outcomes and make decisions quicker than any human being could. But these same tools also present challenges when it comes to ethics because they are capable of doing things we cannot anticipate or control.

Fortunately, the machine learning used to build new artificial intelligence models and the resulting interest in ethical use are both combined in an emerging field with many experts working on solutions. One way to avoid unintentional biases is by using a technique called “generative adversarial networks” or GANs. This system creates two AI agents that compete with each other, and the hope is that they will learn from one another without any bias. They are used widely in image generation, video generation and voice generation.

generative adversarial networks” or GANs presented with digital boxes

Options for allaying fears about loss of control or redundancy

Many people and AI experts are concerned about AI technology. These fears seem to stem from a few common root causes. These include general anxiety about machine intelligence, the fear of mass unemployment, concerns about super-intelligence, putting the power of AI into the wrong people’s hands resulting in abuses, and general concern and caution when it comes to new technology.

Indeed, we all want to do the right thing. It is accepted that an ethical framework could shape a system of moral principles that helps us live up to our own ideals and values, as well as those in society.

AI still remains today one of the most fundamental transformative technologies that we have ever seen in the history of mankind and with this transformative power comes fears about its long term potential.

Mass of people representing mass unemployment

The three major areas of ethical concern

Artificial intelligence presents three major areas of ethical concern and ethical issues for society: privacy and surveillance, bias and discrimination, and perhaps the deepest, most difficult philosophical question of the era, the role of human judgment, said Sandel, who teaches a course in the moral, social, and political implications of new technologies.

  • Privacy and surveillance:

Most AI experts agree that this is the area of most concern for ethics. The more we are part of current technological progress and data is used to make informed decisions, the more data is collected by AI algorithms about us which results in ethical data privacy issues. Data can be incredibly sensitive and it’s important to have a say in how it should be used because we don’t want our lives to be exposed out there to drive decisions without our informed consent and permission.

  • Bias and discrimination:

Bias is the act of discarding, judging or treating a person harshly based on their race, gender, beliefs and so forth. The whole problem we are all trying to address around diversity, equity and inclusions as well as human rights.

In ethics it is used to describe any kind of prejudice for or against one thing over another. When we look at biases within the context of intelligent and autonomous systems, this means looking for explainable AI and trustworthy AI that are are not utilising false positives, fake content, impacting human dignity or leveraging human weaknesses to make next best recommendations.

  • Role of human judgment:

Some people say this ethics debate has come about because they are afraid that artificial intelligence will replace them in their jobs, or that ethics will be a job. Others believe we are too focused on how AI can transform our lives for the better and not enough on what it may take away from us. Look at the automation AI offers in the way social media, tweets and blogs are run. Great for automation. We still need someone to evaluate the accuracy of the data to ensure that things make sense for the recipients.

Dealing with these three areas of ethical concern is an important issue which must be dealt with now before society reaches a point where people feel they have no say in how data sets are being used to support a variety of activities in the real world.

AI ethics is the set of moral principles that guide us when we program machines to make decisions or take actions without human input and oversight. The ethics debate for AI has been going on since its inception, but it’s a discussion worth having now because many people feel they have no control over the ethics topic anymore.

 

Gentleman wearing VR set

 

So what should you consider in your Ethical AI building quest?

As a start, make sure to always assess the validity of your internal privacy, security and surveillance but also how data is used to determine actions and implications of profiling on human rights.

If you are building an AI system that will be collecting data about people, ethical principles requires you to inform them of what’s being collected from them. You can also give users the opportunity to consent or not before they provide any information. This is a requirement of the digital infrastructure being built in many countries including Canada, France and more.

Eliminate biases by ensuring your expert teams deploy the right level of empathy, and allow them to talk and share what they find that is disturbing. We have all learned from the recent Google ethical issue cases.

You may also want to use ethics-based design for your machine learning algorithms, which relates to designing a system with some of the strong AI tools in mind. For instance, complying with ethical AI compliance principles can pose challenges for businesses because there’s no clear definition of what it means or how to go about achieving it yet. Some examples include the need to invest in ethics education and ethics training, invest in a robust ethics framework, put together an ethics leadership team to deal with the challenges that come up when developing artificial intelligence algorithms and evolving AI best practices and principles.

If you want to know more:

 

Discover more on my thinking via LinkedInTwitterYouTube, and Slideshare

 

Photos by NeONBRANDHarley-DavidsonJp Valery, fabio, mauro mora, Lux Interaction on Unsplash

Share this post

Menu