Artificial intelligence is making leaps and bounds—we know this. The accomplishments from OpenAI’s and Google’s cascading releases have become so commonplace—they smash through benchmarks, they get high scores on the bar exam—that new milestones no longer seem worthy of much hoopla. Last month, Anthropic, the maker of blockbuster chatbot Claude, was able to grab attention only by telling the world that its new model was too dangerous to release to the public. Now, of course, they’re working on releasing a version of it anyway. On to the next.
At the same time, something is changing in the design industry around AI, but it has very little to do with technological breakthroughs. Simply, designers are actually using it now—or if they’re not, they think they should be. Polls show that adoption has tripled since 2023. AI coaches are in high demand. Instagram feeds are littered with try-this-cool-trick tutorials, and design schools are teaching AI ethics. The rubber is meeting the road.
We’re in a new moment, and it’s a messy one. No one knows exactly how to use AI in their firm or exactly what tools are the “best”—to say nothing of the ethics of it all. Because the technology itself and the culture around it are changing rapidly, this will not be a time of simple answers. But the right questions to ask are getting clearer. Here are five of them.
What do you actually want AI to do?
This seems simple, but it’s not. During a workshop at the recent Design Intelligence summit, several designers said they’d love to train an AI chatbot on every scrap of data from every single project they had ever done. It’s a cool idea, but it was genuinely difficult to take the logical next step: Once you’ve fed a chatbot all of that information, what’s the first question you’d ask it?
The unrelenting hype around AI can create a unique kind of anxiety. It’s easy to fall into the trap of believing that chatbots (and agents like AI personal assistants, which integrate with other software systems to perform tasks with little to no human supervision) can do anything and everything, and that your peers are already racing ahead by AI-ing better than you. It’s not true. Everyone is figuring it out as they go, and the technology has significant limitations—it can’t run your firm for you.
But it can certainly help with some things, and it’s worth dialing in on what they might be. Some designers may find it easiest to begin with the parts of their job they find boring or inefficient and exploring whether AI can take these on. Others will find more success by tinkering with tools and seeing where it takes them. Either approach can work. The point is to move beyond nodding at empty generalities—“AI will change everything”—and start looking at specifics.
What’s your policy?
This is a moment of private dabbling, and of scattershot usage of AI in design firms: The principal is using ChatGPT to write work emails; the procurement manager thinks no one in the firm uses AI; the junior designer experiments with Nano Banana for renderings; the intern runs everything they do through a private Claude account. Now is the time to take a beat to find out how the technology is being used in your firm already, and to set a policy around future usage, even if it’s somewhat loose.
There are essentially two philosophies. One is the top-down approach: Pick one platform and give an account to everyone in your firm—along with instructions on what kind of tasks AI can and should be used for. This is organized, economically efficient, and will give you the clearest sense of what’s working and what isn’t. However, it can also run the risk of stifling creativity.
The other approach is bottom-up: Whoever in your firm is interested in AI gets a reasonable stipend to play around with whatever tools they’re interested in. You periodically get together to share what’s working and what isn’t, and iterate from there. This approach is messy and potentially wasteful, but you may end up with more interesting uses for the technology.
Neither is “right.” But likely, one is right for you.
How will you handle privacy?
The best way to find genuinely helpful uses for AI is to experiment, using your own work as a test case. The problem is that designers work with a lot of sensitive data, ranging from addresses and phone numbers to floor plans and personal communication. By uploading any of it to an AI chatbot, you’re creating a potential security or legal risk.
There’s a lot to unpack around privacy risks with chatbots. In general, AI companies take legitimate steps toward stripping personally identifiable information (PII), like email addresses, from whatever data they ingest. Some say they will not use your input to train their AI models at all; others only extend that courtesy to paid users. It’s worth reading the fine print, but amid a dynamic landscape, it’s safe to assume two things. One, the big AI companies claim to protect their users’ sensitive personal information. Two, that might not be good enough for your clients.
This is not a new problem, and different firms are handling it in different ways. Some will work with an IT consultant to create a closed internal AI system, preventing sensitive information from leaving your own network (this is not as complicated as it sounds, but it takes time and money). Others will preemptively strip PII from anything uploaded to an AI engine.
The technology, law and culture around this issue will likely evolve quickly. But for the time being, it’s worth looking closely at whatever NDAs a client asks you to sign, revisiting your own contract, and checking in with clients directly: “If I upload a picture of your custom closet to ChatGPT, is that a problem?”
How will you work with AI-obsessed clients?
It was once an oddity for clients to bring ChatGPT into the conversation. Increasingly, it’s an everyday occurrence. If your clients aren’t already using AI, they will be soon. In many ways, this is simply the advent of the internet all over again: More people have more access to more information (try typing “How do interior designers mark up product?” into any chatbot for a preview of the kind of information your clients have at their fingertips).
However, AI has its own oddities. Chatbots tend to be sycophantic, finding ways to agree with whoever is prompting them, which can give AI-obsessed clients a false sense of certainty when they bring an opinion to the table. Crucially, AI also changes the dynamic around renderings. Before, designers essentially controlled the flow of visual representation of the project—if you didn’t make or commission renderings, they didn’t happen. Now, clients can spin up endless images on their own, creating a different kind of back-and-forth.
These developments don’t have to be negative—sometimes AI can help a client understand the nuances of a project and come to decisions more easily, or quickly visualize a hard-to-understand concept. But whether it’s a problem or not, it will be a thing. If you haven’t thought about how to communicate with clients when AI is in the loop, start now.
How do you (and your team) feel about AI in your workflow?
It’s easy to talk about AI in relatively dispassionate terms—here’s what it’s good at, here’s a new tool to try—while ignoring the broader cultural context. But artificial intelligence is not a dry academic subject: Depending on the person and the day, it can inspire anything from excitement or euphoria to anxiety and outright dread. While Silicon Valley spends billions to develop the technology, polls show that a majority of Americans have negative-to-neutral feelings on AI. A recent NBC survey found that only 26 percent of the country sees it positively.
The cultural weather around AI will show up in individual design firms too. You may be excited about the potential efficiencies; your assistant may hate AI so passionately that he or she refuses to use it. The reverse could be true. Your clients may feel wildly different about AI than you do.
There are no “solutions” here. It’s simply worth remembering that emotions are part of the broader picture, and need to be considered as carefully as any other aspect of AI—if your team is morally opposed to the technology, getting them on board with privacy standards is beside the point.
