The ethics of AI image generation are a pressing concern, encompassing bias, privacy, and responsibility. AI-generated images produced by deep learning models are becoming increasingly indistinguishable from human-captured photographs or designs. While these advancements offer exciting creative possibilities, they also introduce ethical considerations that need attention.
Ethics Of AI Image Generation
1. A bit of jargon
First, let’s talk about the different types of copyright mentions you can find online.
Under some pictures, you will see instructions, such as “no commercial use, royalty-free”, etc. Let’s see what they mean:
Commercial use: The image can be used with something that you’ll make a profit from.
Royalty-free: no need to pay fees for using this image.
Requires attribution: you must credit the creator. Sometimes, you can modify the image, and sometimes, you can’t.
Public domain: no copyright or trademark
Creative Commons: free license to allow the image to be used by the public (it belongs to “the public” and not to one artist or company).
2. Who owns what?
There are a few questions that can be asked when talking about copyright. First, can a company claim every piece of art generated through its services or website? What if the company is in the US, but you’re in a different country with different laws? Can you create something with an AI tool and then register it as your own?
One thing that can help us understand is the terms of use. The terms of use can be found on any legitimate website that offers a service or product. For OpenAI, it says that whatever ChatGPT generates, thanks to your prompt, you own it.
You own the output, but can you claim ownership legally? No. What this means is that you do not have to pay OpenAI to use the output that ChatGPT gave you. It does not mean you are legally the owner.
If you look at Midjourney’s terms of use, it says you own the assets you create, to the extent possible under current law, but it also says that you let Midjourney use your art, prompts, or anything else you use to create an image. And the terms are different if you use Midjourney on behalf of a company.
There’s an ongoing debate about who owns AI-generated art. Still, until a final resolution or the creation of a new law, disputes over computer-generated art made by software like Stable Diffusion or DALL-E will be handled case by case. What seems to matter to the court is how much human work was involved in creating the art.
As for using AI-generated artwork, there are some unknown risks since they haven’t been tested in court much. The general rule is that you can use someone else’s prompt (if they agree), but you cannot take their work and claim it’s yours.
In the US, the copyright law says only humans can be granted copyrights, so technically, you cannot claim ownership of something created by AI.
The best way to be safe is to carefully read the terms of use for the AI you’re using and understand what you can and cannot do.
3. Biases and ethics
AI models, reliant on data, can perpetuate biases in their training data. For instance, if the training data underrepresents certain demographic groups, the AI model may struggle to create diverse and inclusive images. You might have noticed it if you use Bing Create a lot.
If you ask for a human without specifying their look, Bing Create will give you Caucasian people most of the time. If you’re lucky, you might see other types of ethnicities represented. Why? We can assume that this is not a choice that was made and that it’s just because the AI was mainly trained on pictures or designs featuring Caucasian people.
Most people and text-to-image providers will always say you cannot use AI-generated images for bad purposes. Let’s look at all the possible issues that could arise from AI-generated visuals.
AI image generation poses significant privacy concerns as it can produce highly realistic images of non-existent individuals. This technology could be abused to create deceptive profiles, deepfakes, or explicit content using non-consenting individuals’ likenesses. Could you imagine the trouble if someone used your face to create content you didn’t agree to be in?
Another issue that could happen with AI-generated visuals is increased scams that rely on using someone’s identity. Scammers could easily hide behind AI-generated pictures or videos to gain the trust of their victims.
On Bing Create’s website, Microsoft took the time to explain how they address using AI creation responsibly:
We take our commitment to responsible AI seriously. We are working together with our partner OpenAI, who developed DALL·E 2, to deliver an experience that encourages the responsible use of Image Creator. To that end, we have incorporated OpenAI’s safeguards and additional protections into Image Creator.
To name a few: We have put controls in place that aim to limit the generation of harmful or unsafe images. When our system detects that a potentially harmful image could be generated by a prompt, it blocks the prompt and warns the user. We will allow living artists to report their names to us to limit the creation of images associated with their names.
(…) We encourage users to take care of the prompts they submit and the usage of images they create with Image Creator.
You May Also Like:
- How To Pass AI Detector Tools With ChatGPT
- How To Optimize Your Resume And Cover Letter With ChatGPT
- AI Scriptwriting: Confessions Of A Professional Scriptwriter
Conclusion
While AI-generated images open up exciting creative possibilities, they also raise important ethical questions. Understanding who owns the rights to AI-generated art can be a bit tricky. AI programs can sometimes pick up biases from their training data, which can lead to a lack of diversity in the images they create. This can be a problem, and it’s up to you to make sure you prompt correctly to represent all the colors of humanity.
AI image generation also raises privacy concerns. It can make very realistic images of people who don’t even exist, which can be used for dishonest purposes, like creating fake profiles or misleading content.
To handle these issues responsibly, following the rules and guidelines provided by the AI tools you use is crucial. Some companies are taking steps to make sure their AI tools are used responsibly and don’t create harmful or misleading content. So, when you’re working with AI-generated visuals, always be careful and think about the ethical aspects to avoid potential problems.
