AI Policy

This policy applies to student work in the Experience Design Certificate Program and any of our other educational programs.

TL:DR - Be Curious. Be Rigorous. Be Human.  (Don’t Be Lazy.)

In writing this policy we have a single purpose: making sure that our students and our community of thought are getting the most out of our time together. AI is exciting, terrifying, emergent, and here to stay. Our hope is that our time together results in a better, smarter, more informed, and more intentional relationship with AI for all involved. As with everything we do at Odyssey Works, our aim is for a collective line of enquiry that is both ethical and practical. Faced with this emergent technology, we urge you to ask: Why should we use this? How should we use this? And what are the implications of using this?

We at Odyssey Works have a complicated relationship with AI. On the one hand, AI is, or has the potential to be, a transformative superpower in the very spaces where we currently work. Responsive systems capable of improvisation, rapid iteration, incredible personalization, and accessible prototyping (just for starters) have seemingly limitless potential for immersive, interactive, personalized experience design. Like many of you, we have worked on previously unimaginable projects utilizing AI. Much of our (and our community’s) design and consulting work often involves AI in one form or another.  Ida Benedetto, an Odyssey Works board member, has created a system to use AI personas to help you workshop your work before handing it in (we recommend trying it for your assignments). We’re sure you have your own examples. There is so much possibility; we would be foolish to ignore it. Most of you are working in fields where you’ll have to become accustomed to AI, and where knowing how to use it to your advantage will be necessary as your field transforms. We would be remiss not to look at this in our time together.

On the other hand, there are huge concerns. Odyssey Works wrote and creative directed the TED talk that kicked off the 2025 TED conference. It was by one of the three “Godfathers of AI”: Yoshua Bengio. He came to us because he was terrified that his creation would kill everyone. Working on the talk gave us nightmares.  As we worked with him, it became clear that, though the total annihilation of the human race was his big concern, the thing that hit him in the gut was the ceding of human agency to the machine. This is ours as well.  In this program, our wish is for technology to empower, not disempower, all of us. We want you to embrace these tools, but not at the expense of your creative growth, your exploratory thinking, or your sense of self.  Generative AI has proven to be incredibly useful in the rapid-fire prototyping of the EDCP Spring Semester assignments. We don’t expect you to hand-develop illustrations or take the perfect reference picture (unless that’s your thing).   We also have seen fantastic interactive systems developed in partnership with AI for our assignments. Fantastic. The tools are fast, and we love that you can work with a system to bring your ideas to life.

But there is a difference between using the tool and having it do the work for you. You know you’ve done the latter if your image looks like every other AI. You know you’ve handed over your agency to AI if you find yourself catering your design to what AI has generated, rather than using AI to support your larger vision. You know you’ve done the latter if, at the last minute, you asked your AI to throw together a few images and some text and you used exactly what it came up with. AI is a tool, and you should no more allow it to design for you than you should allow social media to tell you what to think. To paraphrase Brian Eno (who creates extensively with AI), it does not, in the end, save time, and if it does save time you are either not going deep with it or you are ceding your creativity to it. AI only learns from large amounts of data. Like any algorithm, it simply seeks to understand that data set and push your queries to align with the middle zone of that information. Most, if not all of these data sets also represent a specific world view, replete with biases. Working with the AI means wrestling with it to find the worldview you are seeking to present. 

So what would it mean to have AI enhance your creativity rather than colonize it? This is a question we hold together. This is what we wish to explore. The possibilities could very well be endless. We have seen that it can be used to generate new modes of interaction, quick interesting prototypes, and fun thought experiments. We have seen, too, that there are times when we are stuck, when we can’t come up with the word or the image and it is tempting to give up quickly and ask the machine to do it. It can become second nature to give over to ease rather than experience the joy of a very human struggle. At those times, we take a beat. In that struggle is growth, and is comprehension, and is the vision we seek to cultivate. 

Most academic AI policies are concerned with cheating and with IP. These are worth mentioning, but we are not too worried about it. Go ahead and cheat in this program, should you choose. We don’t wish to have the kind of relationship with you that involves policing your conduct. We do ask you to be transparent about your AI use. Cite it as you would any other work you’ve used. In our programs you can reproduce copyrighted work for assignments under the fair use clauses of the various copyright laws (though not if you then present the work commercially). But good conduct necessitates citing that work nonetheless. Get in the habit of finding the names of the people who make the images and sounds you use. Cite them in your credits if not on the images themselves. It’s a very good practice. Cite the AI you use as well. We prefer the phrasing “created using generative AI” or “created using ChatGPT.” 

Beyond relinquishing your agency, there are environmental and copyright concerns with the use of AI.  Know that if you have a free AI account, what you put into the system becomes part of its training – you lose control of your IP.  As previously mentioned, AI models are trained on mass data sets scraped from the internet, including the data that is fed to the various models. This can be 1) beneficial as you develop a deeper understanding of how to use the tool and it develops a clearer understanding of how to work for you, and 2) puts your unique point of view and creative output at the mercy of the tool without your consent, credit, or compensation. In aggregate, AI also uses a lot of energy. Be conscious of this, yes, but be conscious of it within the full context of how you think about your energy use. Research has shown that, in fact, general text based queries do not use a significant amount of energy. Even making an image does not take a lot of energy. Generating video takes a notable amount of energy, but not a catastrophic amount. A feature length AI film, however, takes a great deal of energy. What is that compared to how much energy a traditional filmmaking process takes? What is lost? Or gained? We recommend thinking about this energy use in the context of all your energy use. Do you drive a car? Ride in planes? Eat meat? Cook with gas? AI is now a part of this calculus; how do you make your choices?

Experience designers have an additional challenge when it comes to AI. What is the experience of life we wish to have with this new technology? Who do we become and how are we shaped by the tools we use? Are the tools working for us or we working for them?

Note: This policy was written without the use of AI.

Published 2026