Springboards takes aim at ChatGPT sameness in new ‘Dangers of AI’ experiment

Springboards is taking aim at what it calls the growing “sameness” of generative AI, launching a creative experiment designed to show how easily large language models such as ChatGPT can collapse distinct ideas into near-identical advertising. 

The creative platform has released a new AI-created piece titled The Dangers of AI, which recreates and reworks ChatGPT’s Road Trip commercial to demonstrate how quickly large-scale generative models can produce polished-looking work that drifts into copyright-sensitive territory and homogenised creative output.

Rather than celebrating what generative AI can make, the experiment deliberately dramatises what happens when many teams rely on the same models, prompts and workflows.

“We’re very aware of the irony here,” said Springboards CEO and co-founder Pip Bingemann.

“We’re dramatising the problem of large models sending everyone to the same place by deliberately using a technique that exposes how easily they drift into infringement. But sometimes the only way to show the danger is to step into it. This work is about making those risks visible, not pretending they don’t exist.” 

Springboards said the experiment began inside its own creative process, where the team explored nearly 20 different ways to describe its brand using large language models, only to find that different prompts repeatedly produced similar outputs.

The team then partnered with an AI production crew to recreate a recent OpenAI ad spot as a test case. The first iteration, according to Springboards, was “uncannily close” to the original, immediately raising questions around actor likeness, copyright, and the speed at which generative systems can slip into unsafe territory without malicious intent.

From there, the team intentionally pulled the work away from recognisable voices and identities, sacrificing perfection in favour of exposing the underlying issue.
What started as a playful experiment, Springboards says, became a broader commentary on three converging risks: LLMs increasingly sounding alike, AI-generated content overwhelming human-created work, and high production values becoming cheap and fast to replicate, even as originality quietly erodes.

CMO and co-founder Amy Tucker said the project reinforced why generic generative models are ill-suited to lead creative thinking.

“This experiment really showed the dual reality,” Tucker said. “The models are powerful, but they narrow creative possibilities as much as they expand them. Creativity needs tools built for the craft, not systems that smooth every idea into the same outcome.

“That’s why at Springboards, we aim to be an enabler, not the final answer. Springboards gives teams the variation and space they need to unlock new creative directions while keeping the taste, judgement and originality human.”

For agencies and brand teams, the company said the work serves as a warning heading into 2026, a year where generative AI is will continue to accelerate.

Springboards said the industry faces a widening gap between what large, general-purpose models can produce and what great advertising actually requires: variation, judgement, cultural nuance and originality. The company’s Experiment Drops series will continue to release monthly projects designed to spark debate about the future of creativity and advertising in the age of AI.