#29 The Top Four Ethical Considerations of AI in Design
12:10 Share

Hello friends and welcome back to the future of UX podcast your go-to source for all things about design technology and the future to future proof your design career and understand what is going on in the design and the tech world my name is Patricia Reiners i'm your host for this podcast i'm working as an innovation designer and consultant in berlin where i'm running my own business

Today's episode is about the ethical considerations of using AI in design. This is such an important topic and I'm going to share with you four things that you need to know when it comes to ethical considerations in artificial intelligence. So I would say let's get started.

Before we dive into the top four ethical considerations of AI in design, let's explore a real-world example that highlights the potential of AI in design. Nike's AI-powered design system, Nike Fit, uses computer vision and machine learning to accurately measure customers' feet and recommend the best shoe size for them.

This is just one of many design projects that are using AI and also, of course, besides the project, there are a lot of design tools that we as designers can implement in our design workflow, thinking about image generation tools like DALI or MidJourney or content creation tools like Jasper, Copy.ai or ChatGPT.

So we as designers of course are so curious about how we can integrate AI in our design process and create more personalized and sustainable solutions for our users. So let's explore the four ethical considerations of using AI in design. Let's get started with the first one which are biases in AI.

AI systems are only as biased as the data they're trained with or the people who are working with this data. So if the training data is biased, the AI system will likely produce biased results. And so designers really need to be aware of these issues and take steps to ensure they're using AI ethically and responsibly. And there's one example I shared that in a podcast a few episodes ago, but also on Instagram.

One of the tools that I'm using on a daily basis is Midjourney. I love Midjourney. It's super great for visualization, for any kind of content. If you need to create a storyboard, if you need any visuals for a website, it's an amazing tool.

But this tool is totally biased from the data that has been used to train mid-journey. One example is if you ask the tool to create an image of a doctor in a hospital.

and a nurse in a hospital. It's super interesting what the tool prepares for you. So Midjourney always prepares four images for you and then you can either like upscale it or do variations on it. And if you ask for an image of a doctor in a hospital, you get four images of white male persons.

which is super crazy. So there is absolutely no diversity. And of course, surprise, it's not true that just like males are doctors, right? So that's not true. There are, of course, also women who are doctors. Oh, that's because of the training data that has been used there.

And the same with nurses, which is pretty interesting, because when you ask for an image of a nurse in a hospital, you only get white females. Super interesting, right? That's also not the true. Of course, there are as well men who are nurses. So this is not like a typical women thing. So it's like there's a lot of diversity. The same with like ethical background, right? So not all the nurses and doctors are white, obviously, right? So there are, yeah, so...

Obviously, that's not the case, so not true. And the reason why the images are created like that is because the data that has been used to train these tools have been like that. So no variety, no diversity. And the people who trained the tool, who kind of like said, yes, that is right with reinforcement learning,

were probably not diverse enough so that they saw these things. So we as designers really need to be aware of these issues and then also really take steps to ensure that we're using AI ethically and responsibly. And when we see these biases, we need to report them.

So this can include using diverse and representative data or training data, testing the AI system for biases and being transparent with users about how AI is being used and how it has been trained. This is something that we are definitely missing. Okay, point number two, privacy in AI.

So AI often collects a large amount of user data, which can raise privacy concerns. This data is stored in the cloud, right? So designers really must ensure that users are aware of what data is being collected and how it's being used. So there needs to be some kind of transparency.

And we as designers or people who are responsible for the business model must also take steps to protect user data from unauthorized access or use. Easier said than done, right? Because like when content is in the cloud, of course, there are some leaks. There are a lot of things happening that you can't really control. So this is definitely a very risky thing.

So balancing the benefits of AI with the potential risks to user privacy is critical. Especially thinking about ChatGPT where you can do basically anything. You can ask anything and a lot of people are using that also for their work, right? So ChatGPT knows so much about you.

and we need to be aware of that. Although there is a new version that you can actually use with JetGPT, I think like kind of like a paid thingy, that you can use it as some kind of like a business tool and then you have more responsibility of how your data is used, which I think is pretty interesting and this is something that we definitely need for all tools. So something we should keep in mind when we are designing for AI, when we are creating AI products. Okay,

standpoint number three intellectual property in ai generated design and this is a huge discussion at the moment so ai generated designs can raise questions about ownership and copyright and we must ensure that they have the legal right to use and distribute ai generated designs right because um

If you ask or if you use, for example, Midjourney or an image generation tool and ask them to create a certain illustration or style with everything or certain elements, it can be Mickey Mouse, for example.

then it's really tricky right because obviously you don't have the right to use mickey mouse in your designs or in your publications you actually need to buy the copyrights for that and if you don't have it you are actually not allowed to use them the same with

the work that you're putting out there in mid journey it's open so people can see what you created and it's not protected or anything so basically everyone can use the same images that are created you can just like find everything online so there is no protection

So yeah, really must ensure that users are aware of who owns the intellectual property rights of AI generated design and ethical design practices are really crucial in ensuring that AI is being used in the best interest of users, right? So that, I mean, we are all creatives here. So I think the worst that could happen to us is if people really steal our work.

and just like copy our work from behind and then put it in an AI that creates new designs based on the things that we created. Things that I would say the design industry hadn't really figured out of how to deal with it, but this is a super important discussion and we need to be aware of it. Like intellectual property is super important. Then point number four, and this is transparency in AI.

And this is definitely a super important topic because we really need to be transparent about how AI is being used in the design process and how the data is stored, what happens with everything. So there's a lot of transparency missing in different parts of design tools. And for us, we really need to inform users of what AI is being used for and how it's impacting the design in the end.

This can help build trust with users and ensure that AI is being used in ethical and responsible manners. A lot of people don't really understand what is going on and what is happening. So being very honest and showing, like being transparent about what is happening, how the data has been trained, how the things are stored, etc.

how AI is used, what AI is actually taking over, will build a lot of trust also with our users. So being transparent can help identify potential ethical issues and allow designers to address them before they become problematic. Let me summarize the four points. Number one is bias in AI.

So this is if the training data is biased, the AI system will likely produce biased results. So we really need to be aware of these issues and take steps to ensure that we are using AI ethically and responsibly. Number two: Privacy in AI. We must ensure that users are aware of what data is being collected and how it's being used. Number three: Intellectual property in AI generated designs.

We need to ensure that we have the legal right to use and distribute AI-generated designs. And number four, the last point, transparency in AI. There, it's about the issue of informing users of what AI is being used for and how it's impacting their design. Super important topics and definitely worth to discuss, dive deeper in and be really open and aware of

also the issues that are happening because this is what will actually make or break the whole AI in design discussion. Thank you so much for listening and if you want to hear more about AI ethics, the future of design,

You can sign up for my free weekly newsletter where I share resources, tools, tips and tricks for designers to keep up to date with the design industry and make sure to know what is going on at the moment. So you can just sign up for free and you'll find the link in the description box. Thank you so much for listening and hear you in the future.