As generative AI becomes woven into creative workflows, a critical question is emerging that goes beyond technical capability: do these tools genuinely empower creators, or do they quietly erode the decision-making that defines artistic mastery? A symposium at Northwestern University brought together researchers, designers, and artists to examine how AI is fundamentally reshaping what it means to be creative across multiple disciplines, and the findings suggest the industry has been asking the wrong questions .

What Does "Creative Agency" Actually Mean When AI Is in the Room?

On March 6, Northwestern University's Center for Human-Computer Interaction and Design (HCI+D) hosted a symposium titled "Creative Agency in the Age of AI" that examined a tension rarely discussed in mainstream coverage: the difference between having access to a powerful tool and maintaining genuine creative control . This distinction matters profoundly because it shapes not just how creators use AI, but whether they feel authentic ownership over their work.

"As AI tools are increasingly shaping creative work, it is critical to both understand how current tools are affecting humans' creative agency and how we can create a future where human creative agency flourishes," said Duri Long, assistant professor of communication studies and computer science at Northwestern University.

Duri Long, Assistant Professor of Communication Studies and Computer Science, Northwestern University

The symposium featured demonstrations of cutting-edge AI applications, including Google DeepMind's Lyria and Magenta music generation models, which allow users to create high-fidelity audio tracks from text or image prompts . But the real focus wasn't on what the technology could do. Instead, panelists and researchers examined how interface design, team dynamics, and workflow integration either support or undermine a creator's sense of control over their artistic choices.

How to Design AI Workflows That Preserve Creative Agency

• Maintain Granular Control Over Decisions: Recognize that AI tools can either amplify your creative vision or replace your decision-making. The key is designing workflows where you remain the decision-maker at critical points, not delegating core artistic choices to automation.
• Preserve the Felt Experience of Control: Panelists emphasized that creators need to maintain a psychological sense of control over their work, even when AI handles technical execution. This means choosing tools with transparent processes and the ability to override or refine AI-generated outputs.
• Build Expertise Before Relying on AI: Researchers discussed how expertise moderates AI reliance. Creators with deep knowledge of their craft can use AI more effectively because they understand what they're asking for and can evaluate results critically, rather than accepting whatever the tool produces.

The Hidden Cost of Removing Friction From Creative Work

One of the most provocative discussions centered on what panelists called the importance of "suffering" through a medium to develop taste and mastery . This isn't romantic nostalgia about the old days. The concern is concrete: if AI tools remove friction from the creative process, do emerging artists lose the opportunity to develop the judgment and intuition that separates competent work from exceptional work?

Ethan Manilow, a senior research scientist at Google DeepMind and Northwestern PhD graduate, demonstrated Lyria and Magenta during the symposium's opening demo workshop . But even as he showcased the technology's capabilities, the broader conversation acknowledged a real risk: without intentional design, AI could lead to artistic homogenization, where creators gravitate toward similar outputs because the tools themselves encode certain aesthetic preferences.

"The demos showcased cutting-edge tools and artwork that leverage AI to support rather than augment human creativity," noted Duri Long.

Duri Long, Assistant Professor of Communication Studies and Computer Science, Northwestern University

How AI Reshapes Team Dynamics and Attribution in Creative Work

A second major theme emerged around how AI fundamentally changes creative teamwork. When a music producer uses generative AI to create a baseline track that a human collaborator then refines, who gets credit? How does that affect team morale and collaboration? Panelists highlighted both opportunities and risks: AI can enhance coordination and decision-making, but it also introduces potential for social isolation and fear of reputational stigma if creators worry about being perceived as "cheating" by using AI .

This concern is not hypothetical. In creative industries, reputation is currency. If using AI tools becomes stigmatized, creators may conceal their use, which undermines transparency and makes it harder for teams to develop shared norms around AI integration . The result is a kind of shadow adoption where AI is used but not discussed, preventing the industry from developing healthy practices around human-AI collaboration.

What Research Priorities Emerged From the Symposium?

The symposium concluded with small-group brainstorming sessions where participants developed mini-research proposals on future directions . The topics reveal where researchers believe the real work needs to happen:

• AI as Social Glue: Using AI as a tool to foster creative teamwork and coordination rather than as a replacement for human collaboration.
• Expertise as a Moderator: Exploring what role expertise plays in determining whether creators become overly reliant on AI or use it strategically to enhance their existing skills.
• Productive Discomfort: Designing AI systems that push creators out of their comfort zones rather than enabling them to stay in safe, familiar territory.

"Our hope is that this symposium paves the way for future interdisciplinary research collaborations, leading to a future where AI supports human creative agency rather than undermining it," said Duri Long.

Duri Long, Assistant Professor of Communication Studies and Computer Science, Northwestern University

The implications for the broader creative technology landscape are significant. The technology itself is not the problem. The problem is how it's designed, integrated into workflows, and normalized within creative communities. A generative AI tool that gives creators granular control and maintains transparency about how decisions are made will support creative agency. One that obscures its decision-making or removes human choice points will undermine it, regardless of how powerful the underlying AI model is .

As AI becomes mainstream in creative work, the conversation needs to shift from "Can AI create?" to "How do we design AI tools that keep humans in control of their creative vision?" That's the question Northwestern researchers are now working to answer through interdisciplinary collaboration between engineering, communication, and design.