The Quest for an AI-Crafted Super Bowl Ad
When filmmaker and “m ss ng p eces” founder Ari Kuschnir set out to create a Super Bowl-style ad using only AI tools, he envisioned a future where storytelling could bypass casting calls, filming schedules, and budget constraints. Super Bowl commercials, known for their high stakes, emotional resonance, and cultural impact, are a $7 million-per-30-second investment for brands. Kuschnir’s experiment aimed to test whether AI could replicate—or even reinvent—the magic of a big-budget commercial using text prompts alone.
Armed with ChatGPT for ideation and Google Veo 2 for video generation, Kuschnir’s goal was twofold: to explore AI’s potential for democratizing ad production and to probe its limitations. The result? A surreal, choppy, yet oddly fascinating mashup that reveals both the promise and pitfalls of relying on AI for creative storytelling. This experiment, while branded a “failure” by Kuschnir, offers critical insights into the evolving relationship between technology and human creativity.
The Role of Google Veo 2: Realism Without the Reel
Google’s Veo 2 stands at the forefront of text-to-video AI, generating 10-second clips with striking realism. Unlike earlier tools that produced jittery or abstract visuals, Veo 2 leverages advanced diffusion models to render lifelike physics, lighting, and textures. For example, a prompt like “a pickup truck speeding through a desert with fireworks exploding overhead” yields a cinematic sequence complete with dust trails, shimmering heat waves, and sparks cascading against a twilight sky.
Yet its limitations are glaring. Veo 2 lacks audio support, struggles with character consistency (e.g., a cowboy’s face might morph between clips), and offers no iterative control—users can’t tweak outputs beyond refining their initial prompt. For Kuschnir, this meant stitching together disjointed clips of AI-generated cowboys, eagles, and explosions—each visually impressive but narratively fragmented. “You get these beautiful moments,” he explains, “but no way to string them into a coherent story without heavy post-production.”
Comparatively, tools like OpenAI’s Sora promise longer, more dynamic videos, but Veo 2’s strength lies in its photorealism. However, as Kuschnir discovered, realism alone doesn’t equate to storytelling. The absence of audio—a core component of Super Bowl ads’ emotional pull—left the project feeling hollow, forcing him to rely on external tools to inject personality.
Challenges of Originality in AI-Generated Content
The experiment’s first hurdle lay in ideation. ChatGPT, tasked with brainstorming “2025-style” Super Bowl ad concepts, regurgitated familiar tropes: patriotic montages, rugged trucks, and earnest Americana. When Kuschnir pushed for originality, the AI suggested absurd mashups like “a robot cowboy herding electric sheep”—a concept too abstract for Veo 2 to visualize coherently.
This highlights a core issue: AI models like ChatGPT and Veo 2 are trained on existing data, making them adept at remixing trends but incapable of true innovation. “The AI can mimic the style of a Super Bowl ad—slow-motion shots, aspirational voiceovers—but it doesn’t understand cultural nuance or subtext,” says Kuschnir. For instance, ChatGPT proposed a spot celebrating “American resilience,” leaning on stock imagery of eagles and flags, while Veo 2 rendered these ideas with technical precision but no emotional depth.
The paradox is clear: AI excels at emulating existing styles but falters at inventing fresh narratives. Kuschnir’s experiment became a “surreal version of something we’ve already seen,” highlighting the gap between algorithmic replication and human ingenuity.
Patching the Gaps: Sound, Music, and Editing
Faced with Veo 2’s silent, erratic clips, Kuschnir turned to traditional tools to salvage the project. In Adobe Premiere, he layered voiceovers from ElevenLabs, a twangy AI-generated country-rock track from Suno.ai, and sound effects to inject rhythm into the chaos. The ad’s accidental quirks—a robotic voice declaring “Ame-re-rica” or a pixelated eagle glitching mid-flight—became its charm.
Sound design played a pivotal role in masking AI’s shortcomings. For example, a clip of a truck swerving through mud felt lifeless until Kuschnir added tire screeches, splattering effects, and a rumbling bassline. Similarly, Suno.ai’s off-key anthem, which mangled lyrics about “freedom’s fiery soul,” unintentionally parodied the earnestness of real Super Bowl ads. “The mistakes made it interesting,” Kuschnir admits. “They gave it a weird, human edge.”
This post-production phase consumed over 80% of the project’s 8-hour timeline. While AI generated the raw visuals in minutes, the human labor of editing, soundscaping, and tonal adjustment proved irreplaceable.
Audience Reactions: The Uncanny Valley of AI Ads
The video’s comments section split into two camps. Critics called it “choppy,” “dull,” and “unconvincing,” citing the uncanny valley effect of AI’s near-perfect-but-slightly-off visuals. One viewer compared it to “a child’s mimicry,” lacking conceptual depth. Another lamented, “I was bored after 20 seconds—not because of the content, but because it felt emotionally flat.”
Others, however, praised its accidental surrealism. “The odd accidents are magical,” wrote a commenter, arguing that brands should lean into AI’s quirks before hyper-polished tools erase them. A filmmaker noted, “The glitches remind me of early CGI—awkward but full of potential.”
The debate underscores a key tension: Should AI streamline creativity or preserve its messy humanity? Critics warn of a future where “data-driven briefs” replace creative risk-taking, flooding media with formulaic content. Kuschnir’s experiment, with its jarring cuts and absurdist tone, inadvertently became a critique of AI’s current role—a tool that democratizes production but risks homogenizing artistry.
The Broader Implications for AI in Creative Industries
Kuschnir’s project mirrors broader industry trends. Brands like Nestlé and CarMax already use AI for localized ads, while platforms like TikTok integrate generative tools for user content. However, Super Bowl ads represent a unique challenge: they demand cultural relevance, emotional resonance, and novelty—qualities AI struggles to synthesize.
For instance, iconic ads like Apple’s “1984” or Volkswagen’s “The Force” succeeded by subverting expectations, a feat requiring human intuition. AI, by contrast, defaults to proven formulas. As Kuschnir observes, “ChatGPT gave me tropes. Veo 2 gave me clips. But neither understood why those tropes work or how to reinvent them.”
Ethical concerns also loom. If brands rely on AI for ideation, could it perpetuate stereotypes or erase diverse perspectives? One commenter warned, “A world where content is ‘what someone typed in’ is a world without authentic voices.”
AI as a Collaborator, Not a Creator
Kuschnir’s “failed” experiment offers a critical lesson: AI like Veo 2 democratizes production but cannot yet replace human intuition. While it eliminates logistical barriers—no need for crews, locations, or actors—the lack of originality and emotional resonance in AI-generated content reveals its role as a collaborator, not a creator.
The project’s accidental success lay in its imperfections. By embracing AI’s “weirdness,” Kuschnir created a self-aware parody that critiques both Super Bowl ads and AI’s limitations. For now, the most compelling stories will emerge from artists who harness AI’s efficiency while infusing projects with their own vision—flaws, weirdness, and all.
As Kuschnir puts it, “AI is a brush, not the painter.” Tools like Veo 2 and ChatGPT can sketch the outline, but humanity must fill in the color. The future of creativity isn’t about replacing artists with algorithms—it’s about empowering artists to reimagine what’s possible.