Year:
2026
Focus:
AI Design
For:
Adobe x Parsons
OVERVIEW
Not Generated is a spring 2026 collaboration between Adobe and Parsons School of Design at The New School to examine the evolving relationship between creativity and technology. The initiative explores how designers and creators position generative AI within their process, and how questions of authorship, responsibility, and intention inform creative decision-making.
As one of twenty students selected to participate in the Creative AI Lab, I speculated on how the future of creativity could be consciously shaped through responsible experimentation and human vision. The tools that exist today are oftentimes built for creators and platforms, not for the people consuming content.
PROBLEM
How might we give vulnerable populations (elderly, young learners, new users without trained eye for AI) a real time way to understand and verify AI-generated content, especially when trust and privacy are essential?
AFFINITY MAPPING
Across all research, the same fracture kept surfacing: the people most affected by synthetic media are the least equipped to navigate it.
TARGET USER
To uncover the challenges of the people most susceptible to falling behind, a persona was created to better understand their journey.
LOW-PROTOTYPING
Iterating on potential designs
Considering target users want all the necessary information in plain language to be informed, iterations have to include verdict before score, no more than one decision at a time, no false certainty, and community action over individual judgment.
THE VISION
A real-time content trust layer for the people who need it most.
Scanner
NOTIFAI reads C2PA Content Credentials locally and surfaces a plain-language verdict on your paired phone, with thresholds and flag language tailored to what you're actually looking at: news, medical, art, professional identity, or dating.
Threshold
Where current platform labels remain high-level and creator tools focus strictly on publishing, NOTIFAI decodes raw C2PA metadata directly for the viewer by evaluating the exact ratio of human footage to labeled AI graphics, institutional anchoring, human sign-off status, and contextual completeness.
Community Trust Hub
The community hub receives only what you choose to share, connecting your result to family members, verified institutions, and a new class of human verifiers trained to review flagged content, all without logging any behavior.
USER TESTING
Most AI transparency tools were built to help platforms comply with regulation not to help people navigate an information environment that wasn't designed for them.
NOTIFAI inverts that: the scan points at content and trust is earned through disclosure rather than penalizing AI use outright, making it a tool creators and institutions can actually work with.
Though some risks to consider include:
OUTCOME
By 2036, synthetic content will be everywhere.
And that framing gave me room to ask questions present-tense design forecloses. What does it mean to prove something is real? Who gets that proof, and on whose terms? The technology is ahead of the governance, and the governance is ahead of the interface. Some questions moving forward that I’ll continue to ask are:
