Using AI Can Be Ethically Iffy. Here's How to Do It Right (2025)

Like it or not, artificial intelligence is here to stay. You’ve probably heard a lot of chatter about how it’s an incredible tool that can transform society and a malevolent force that steals work and is destined to put us all out of a job. The truth, of course, lies somewhere in between.

As with any tool, the nature of AI depends on the people designing and using it. What distinguishes AI from many other resources available in the modern businessperson or student’s toolbox is its incredibly high ceiling, which current models haven’t begun to challenge.That said, the landscape is broad even now, and the proliferation of options can be confusing and intimidating.

How do you leverage AI to simplify your life without wandering into morally ambiguous territory? Here are a few tips that we recommend.

Do Your Homework

Before you use AI to help with your homework, you need to do some homework on the AI tools you choose. The reality is that visibility into how large language models and other AI systems are built and trained is mainly at the discretion of the people creating them. It doesn’t help that regulations around AI are still very much in their infancy—unsurprising given the rapid evolution of AI technology.

To identify ethical tools, you must first understand how the AI has been trained. One of the industry's most widely discussed topics is the disregard for copyright or content ownership. Some AI models are trained on copyrighted material and use it for output, often amalgamating several sources without any credit (or compensation) to the original authors of that content. Several lawsuits are currently ongoing.

Another issue facing AI is the misleading or false output it produces. AI is notoriously bad at distinguishing between low- and high-quality sources of information. It often gets polluted with deceptive or outright false information, which it then outputs as fact. As my colleague Neil J. Rubenking has written, AI is the nemesis of truth.

That said, it’s still possible to seek out companies and tools committed to the ethical use of AI. Look for investments in ethical usage that go beyond lip service, like companies that hire employees to monitor copyright violations or inaccurate and deceptive output. AI devs that employ an ethics department are usually happy to advertise it on their website or in the press.

As much as possible, you should also look at who’s building and funding the AI tools you want to use. Is the company behind those tools taking large infusions of cash from tech giants with shady reputations, or is it a team that seems genuinely interested in how AI can be a boon for humanity? While you can never truly parse a company’s intentions online, you can often get a general sense of its operations through how it discusses its products.

Run Your Own Tests

To understand how accurate an AI's output can be, put it through the wringer before you start using it for work that matters. Ask it complex questions that you know the answer to, or request complex outputs and see what the model produces. This can give you a sense of the tool's general accuracy and also how sophisticated the model behind it is. Pay attention to how the AI responds to corrections or identifies potential issues with its responses.

(Credit: PCMag / Google)

You should also feel free to pose ethical questions. If you’re using a chatbot to help with homework or research, ask how it was trained, how it ensures its accurate responses, and what steps it takes to avoid sharing fake info. You can even ask about the company directly, giving insight into how it prioritizes things like fair use of copyrighted material versus outright piracy.

(Credit: PCMag / OpenAI)

Always Check Your Sources

Anytime you’re using AI for factual purposes, check its sources. Before you use any data, ask the model to provide (multiple) sources for the information it's offering. You can double-check it in other ways, either by searching for that information online or even by asking another AI. For example, if you're using Google's AI Overview feature, click the link icon next to each item to see where that information came from and verify its accuracy.

(Credit: PCMag / Google)

It’s also good practice to source images and art to ensure you’re not using copyrighted material. For instance, some image generators are only trained on open-source and non-copyrighted material. You can also run any image through a reverse image search to ensure you’re not using work explicitly derivative of an existing artist.

It’s thorny ground, both legally and ethically, but at the very least, you can take the time to ensure you’re personally comfortable with any AI resources you end up using in your own work. It’s also vital to be transparent; any time you use AI-generated content for work or school, disclose that fact to your client, employer, or teacher.

(Credit: Mitsua Diffusion One)

If you’re looking for a chatbot, consider Anthropic's Claude. The Anthropic team broke off from OpenAI and is well-known for its “constitutional AI” framework, which is designed to make AI safe, ethical, and more helpful to real people. Creating a Claude account requires agreeing to a usage policy that prohibits using Claude for harm, like producing violent, abusive, or deceptive content. Note that Anthropic has received significant investments from Amazon and Google.

(Credit: Anthropic)

Mitsua Diffusion One is a free text-to-image generator trained on non-copyrighted material and built around a simple and intuitive interface. For businesses or individuals looking for good document management tools, DocuWare has expressed a commitment to accuracy, safety, and transparency in deploying AI in its software.

(Credit: DocuWare)

Regardless of what tools you choose to use, taking a few moments to educate yourself about their provenance and processes can not only make you feel better about their safety and ethics, but can make using them easier and more efficient.