AI Ethics

Acceptable Uses of AI

With personal usage, you can tap into AI in whatever way you feel comfortable with. When it comes to schoolwork or tasks for a paid job, it gets trickier. The use of AI may be perfectly okay in one context and unacceptable in another. The baseline you can assume for almost every situation is simple: if you used AI, you should acknowledge it. Problems occur when you claim to do something on your own when, in fact, you used AI in the process. No matter how much or how little you use AI for a project, it’s always best to disclose it.

Many teachers and employers accept or encourage the use of AI as long as it’s disclosed or used in agreed-upon ways, while some don’t. For example, one teacher may allow the use of AI to outline a paper and then review it after it’s written. Another teacher may ban the use of AI altogether. One isn’t inherently better than the other; it’s all about matching expectations and following guidelines.

The baseline you can assume for almost every situation is simple: if you used AI, you should acknowledge it.

Using AI when it has been explicitly forbidden, especially in a school setting, is cheating. Cheating with AI can come with the same consequences as any kind of cheating, including failing a project or paper, detention, or even expulsion. You may think you can “get away with it” or that you can use AI in a way that won’t be caught, and every once in a while, that may happen, but the risks of cheating outweigh the benefits. On top of that, you’re robbing yourself of the opportunity to learn the way the teacher intended. If you outsource all of your thinking and planning to an AI, you don’t get to develop those valuable skills.

At the end of the day, you need to make sure you understand what’s expected of you and act accordingly.

“Theft” and Job Loss

AI is trained on and utilizes human-made data, often without the creator’s consent or knowledge. Many people consider this theft of the creator’s work. It becomes even more complicated when AI starts to take the place of human workers, particularly when the output of those workers was used to train the AI model.

Let’s look at an example of how this could play out: before AI, if you wanted to give a custom illustration to a friend, you might hire an artist to make it. After you paid, you had to wait for it to be done. Now, you could put the same illustration request into a generative AI model, even asking it to emulate the particular artist’s style, and receive an illustration nearly instantly for free or a small fee. You still get the illustration, but the artist also isn’t getting paid for a style that originated with them.

The way that money is changing hands is evolving, and we don’t know the full impact of that shift yet.

You can see the economic impact of this kind of shift at a larger scale when employers lay off workers because their tasks can now be done cheaper and faster by AI. The way that money is changing hands is evolving, and we don’t know the full impact of that shift yet.

It’s possible that the way AI models access data to train on may shift in the future. For now, this type of training is legal in the U.S.

Environmental Impacts

The biggest environmental issues focus mainly on the data centers required to make AI possible. These large buildings house computers that give off a lot of heat. Millions of gallons of water are used to keep the systems from overheating. These data centers also require massive amounts of electricity. This can hugely impact local freshwater storage and electrical grids, strain nearby ecologies, and impact the people who live nearby with higher prices, difficulty utilizing resources, and more. You can learn more about AI’s environmental impact and what’s being done about it on the United Nations’ website.

While there are advancements being made to help cut down on these impacts, it isn’t currently possible to get rid of them completely. Countries around the world are currently battling with how to regulate these impacts, and requirements are changing all the time.

Privacy

As has already been established, AI works because it ingests and regurgitates or edits information from online. As of right now, there is no way to definitively stop information from being scraped by AI. That means anything you or anyone else puts online has a high likelihood of being used to train one or multiple AI models. That includes things like pictures of yourself and your family, your name, private conversations, your email address and phone number, and anything else available online.

The main issue here is that it takes sensitive information out of the hands of the people who own it. From there, it’s possible the information can then be hacked, shared, or sold from the AI model.

AI is an extremely powerful and impactful tool. It’s already changed much about how people and businesses accomplish tasks, and that shift has only just begun. It’s important to understand the ethical issues surrounding AI so that you can decide how and when it’s most appropriate to use it.