We already know designers are using artificial intelligence—for administrative and operational tasks, for visualization and rendering purposes, and sometimes for gathering inspiration and sparking creativity. Industry surveys show that AI adoption has doubled year over year since 2023, suggesting that by next year, nearly all designers will be using the technology.

But learning on the fly can make it difficult to spot the big-picture issues at play, let alone determine your firm’s stance on each of them. And the breakneck speed at which AI is evolving means designers are constantly having to adapt in real time. While many have found themselves swept up in the “Adopt now, ask questions later” mantra, a growing number of designers are beginning to take a more cautious approach to using AI, setting boundaries and establishing codes of conduct for themselves and their firms. Here are the main issues designers face when incorporating the technology into their daily practices, and just some of the ways they are setting guardrails in response.

Privacy
Most design firms take precautions when it comes to handling personal information—both clients’ and their own—in the physical world, and put enhanced cybersecurity protections in place to ensure privacy in digital communications and financial transactions. But many designers, encouraged by the time- and cost-saving benefits, are feeding sensitive material into chatbots without a second thought.

“I’ve had designers tell me they put their QuickBooks into these large language models, and that’s a problem,” says Jenna Gaidusek, a Charleston-based designer and the founder of the consultancy AI for Interior Designers, who has been grappling with these issues since she began teaching her ethics-based certificate program in 2023. “If you’re using a free version, those conversations are never private, because we don’t know what they’re using for training.” There may also be downsides for designers who have upgraded to paid versions of the chatbots. “You don’t know what can be hacked, and to be honest, I don’t trust what they’re saying about how they’re not using that information to train these models,” she explains. “Personal information could always be exposed.”

She says it’s important for designers to remove any identifying information from documents before feeding them into AI models. Joe Carline, partner at New York–based Kligerman Architecture & Design, takes that concept one step further. “We have a strict ban on uploading client-specific data into public AI models to protect client confidentiality,” he says. That includes floor plans, addresses, budgets and private correspondence.

Intellectual Property
Because generative models like ChatGPT rely on data scraped from across the internet—often including work by designers and photographers that is uncredited and used without their consent—Carline says that the biggest ethical question his firm faces relates to the protection of intellectual property and creative integrity. For that reason, he and his partners have developed a policy they call the Human Start and Human Finish Rule, which states, “No AI-generated image leaves the office as a final deliverable.” Though AI can be used as a brainstorming tool, he says, “the final output must be drawn, modeled or vetted by a human designer to ensure that it is buildable and accurate.”

Gaidusek avoids copyright-infringement tanglements by using solely her own images—that is to say, images she owns the rights to—in her prompts. “I don’t take photography from off the internet at all,” she says. “Not even from Instagram or Pinterest.”

Transparency
Many firms have strict guidelines about using AI for administrative and operational tasks only—anything creative is off limits. But for those who use the technology for rendering or brainstorming, several designers noted that it was important that clients have clarity around what’s being presented to them, and that designers not pass off chatbot-generated work as their own. “We have to be careful,” Carline explains. “We want to ensure that the work [we do] is authentically ours. We believe in human-led design, where AI is a tool in the hand, not the hand itself. If an image in a mood board is AI-generated, it must be labeled.”

Brooklyn-based designer Nasozi Kakembo agrees on the need for proper citations. “If we use renderings, we have to be really clear in the accompanying text that these are AI-generated, or identify the components that are not genuine or authentic. That’s not new, because you could do a lot of this stuff in Photoshop,” she says. Yet Kakembo argues that the ease of crafting images with AI tools gives designers a false sense of ownership over their creations.

She tells the story of one designer who digitally altered the image of a room, adding furnishings and accessories to the image that weren’t really there and passing it off as a real-world, completed project. “If the work is not completed or in existence, you can’t claim that it is,” she says. “That’s not ethical. That’s crossing a line.”

Authenticity
Blurred lines about the ethical controversies of using AI are something interior designer Ellen Fisher confronts daily. As the vice president for academic affairs and dean of the New York School of Interior Design, she has a front-row seat for how the next generation of designers relates to AI. “Designers have always been assembling inspirational images and been inspired by the work of others. Now, AI facilitates all of that, but the danger is losing your own original voice,” she says. “AI isn’t just searching for flat images—these models are also creating 3D images. Some students can’t even use CAD yet, but they’re feeding data into these models and saying, ‘This is my idea.’”

These exchanges get at the heart of the nature of creativity itself. “We’re a college. People have to be able to write,” Fisher explains. “If you read enough of these prompts, they all sound the same. We tell students, ‘You don’t have to be perfect. You have to be willing not to be perfect.’ We prize originality over perfection. To be original is something to be protected.”

In the fall, NYSID is updating its long-running Applied Ethics class to include more material on AI, and the school will introduce a new Master of Professional Studies program in digital practice management, led by the award-winning design technologist Anthony Samaha. In the next 18 months, the institution will also be planning a symposium on AI in design and design education. “People are so excited by design that they don’t always think about other equally important aspects,” says Fisher.

Environmental and Community Support
By now, it’s common knowledge that it takes energy and water to power every prompt fed into a chatbot, yet even designers who strive toward sustainable design practices often forget to factor those numbers into their firms’ AI usage equation. Seattle-based designer Jessica Dorling says that when she first began experimenting with generative models, she didn’t think of the hidden environmental costs associated with their implementation. Now she makes a conscious effort to curb her usage.

“The resources that are involved to create—it hurts my heart a bit,” she says. “But a way to combat that is minimizing the number of revisions I do and being more thoughtful of the times that I use it. Now that I’ve gotten over the learning curve and I understand what I’m trying to do, it’s something that I’m thinking about much more than I did.”

Though it’s more for her own efficiency, Kakembo maintains a catalog of successful prompts that she can reference to streamline her iterative processes for future projects. The environmental benefits are just a bonus.

Yet the degradation of the environment is just one lens to consider when it comes to the impacts of artificial intelligence on people. McLean, Virginia–based designer Tracy Morris takes a humanistic approach to chatbot usage because she appreciates working with other humans and prefers to maintain those relationships into the future. “I’m using AI to help with marketing concepts and creating spreadsheets for breakdowns of client invoices and things like that,” she says. “But I’m still going to our renderers and hand-watercolorist. The technology isn’t there yet, but more importantly, I don’t want to lose that connection. I like calling her and hearing about her kids. I like giving her the work. The point is that what we do in design is so high touch. There’s inherent value in that.”