AI is transforming advertising production cycles, compressing them from months to days while slashing costs by over 95%.

Author | Xiaowei

Image source | Shutterstock

AI-generated ads are increasingly making their mark during prime-time sports events.

On June 11, 2025, during the YouTube TV live broadcast of the third game of the NBA Finals, a 30-second ad created entirely by Google's latest video generation model, Veo 3, garnered significant attention.

Despite its absurd plots and exaggerated character movements, which left many viewers scratching their heads, the brand name repeatedly shouted by the characters left a lasting impression.

This isn't the first foray into AI advertising; Coca-Cola and McDonald's experimented with it in 2024. However, the latest breakthrough lies in its astonishing cost-effectiveness: produced on a single computer in just 48 hours at a cost of $2,000, it achieved 100 million views across TV and online platforms, in stark contrast to traditional methods that require hundreds of thousands of dollars and months of production time.

Is this merely a fleeting novelty or does it signify the true dawn of the AI advertising era?

AI Ad Debuts During NBA Finals

Breaking Down the Creation: AI as the "Creative Director," Efficient Output through Human-AI Collaboration

PJ Accetturo, the director of this groundbreaking ad, has a background in traditional film and television but has now transitioned into AIGC content creation. He shared the creative process of this AI ad on platform X, highlighting how AI tools assisted in the core stages:

Step 1: Script Development with ChatGPT

PJ Accetturo first established the ad's overall tone: humorous and eye-catching. This required human input. He then utilized ChatGPT to generate lines based on his requirements, often selecting only a few satisfactory lines from numerous options.

Step 2: Using Gemini for Detailed Storyboards and Text-to-Video Prompts

For instance, "At night, on the streets of Miami, a white man in his 60s, wearing a pink short-sleeved T-shirt and fluorescent green shorts, holds a slightly trembling Chihuahua to his chest, walking while shouting to the camera: 'Indiana has that kind of toughness!'"

Step 3: Inputting Prompts into Google Veo 3 to Generate the Video

Currently, high-quality output from video large models relies heavily on "lottery draws." This ad was produced by selecting 15 clips from a total of 300-400 pieces of generated footage.

Step 4: Editing Footage and Adding Music with Software like CapCut and Adobe Premiere Pro

While six people were involved in this ad, the core creative direction, script, and shots were primarily led by one individual.

Kalshi, the advertiser, is a somewhat controversial financial trading platform. "We are extremely satisfied with the advertising effect so far, and it has garnered significant attention on social media," said Jack Sachs, Kalshi's media representative.

A Market Worth Hundreds of Billions: Low Costs and Short Cycles Drive the Rise of AI Advertising

While AI advertising is still in its nascent stages, customer acceptance has surged rapidly.

According to the "2025 China Advertiser Marketing Trend Survey Report" released by CTR (China Central Television Market Research), 53.1% of advertisers will utilize AIGC technology in creative content, with nearly 20% relying on AI for over 50% of their video creation processes.

Several factors are driving the rapid rise of AI advertising:

First, Extremely Low Costs

Compared to traditional advertising production and distribution models, AI significantly reduces costs at every stage.

A Goldman Sachs research report highlights that AI is primarily applied in five scenarios within the advertising industry: content creation, AI assistants and agents, content recommendation, optimization engines and campaign execution, and modeling conversion. In the context of automatic ad creative generation alone, AI can generate savings of approximately $114 billion.

Alok Sabu, a marketing professor at Georgia State University, believes that artificial intelligence "will lower the barriers to entry for small brands that cannot afford traditional video advertising campaigns."

Second, Rapid Iteration of Video Generation Model Capabilities and Continuous Upgrades in Video Effects

In 2023, the meme video of "Will Smith eating spaghetti" was still widely circulated, and people were skeptical about the reliability of AI-generated videos. However, in less than two years, video generation model capabilities have rapidly evolved.

Take Google's latest video generation model, Veo 3, released in May 2025, as an example. It has achieved breakthroughs in real-time voice-to-lip-sync-to-expression generation, capable of producing natural and smooth lip movements, subtle expression changes, and clear speech based on text content. The deep learning model behind Veo 3, trained on extensive real human speech and lip data, can accurately map speech to lip movements. Currently, Veo 3 is exclusively available to Google AI Ultra users at a monthly fee of $250.

Veo3 Video Demo

Third, Significantly Shorter Advertising Production Cycles

AI compresses advertising production cycles from months to days, enabling the rapid generation of multiple ad versions for market effect verification and optimization. Currently, giants such as TikTok and Meta are already deploying AI in this field.

At the recent Cannes Lions International Festival of Creativity, TikTok launched a new AI advertising feature where users can upload a picture or a short text prompt to generate a 5-second ad video.

Meta plans to fully automate its advertising business with AI by 2026. By then, advertisers will only need to provide product images and budgets, and the Meta platform will automatically generate ad content, match user profiles for delivery, and even allow different users to see different versions of the same ad in real-time.

Controversy Persists: Does AI Advertising Lack Soul?

However, the rise of AI advertising is not without its critics, who focus on audience experience and emotional connection.

Although Coca-Cola's 2024 AI Christmas ad was polished by a professional team, it was still criticized for "lacking soul" and "not being alive." Similarly, the Youtube page of PJ Accetturo, the director of the Kalshi ad, is filled with negative comments like "garbage."

Coca-Cola AI Christmas Ad

At CES 2025, a survey released by consumer intelligence company NielsenIQ also cast doubt on AI advertising.

The survey explored how consumers' brains process AI-generated ads and found that consumers can intuitively identify most AI-generated ads, considering them less attractive than traditional ads. This weakens consumers' overall perception of ads and brands. Even for AI-generated ads deemed "high-quality," the brain memory response they elicit is weaker than that of traditional ads.

This may be because, in triggering deep human emotional resonances such as humor, moving emotions, and shock, current AI ads still have a significant gap compared to traditional top-tier creativity and human actors' performances.

Conclusion

While brands are beginning to place pure AI ads during prime-time, this doesn't mean traditional ad production will disappear overnight.

AI brings not a simple replacement but a profound restructuring. It divides ad production into scalable and automatable parts and those that still require human ingenuity, forcing the industry to reconsider the delicate balance between cost, efficiency, and emotional value.

END This article is an original work of "Intelligent Evolution Theory," welcome to follow us.