Introduction to AI-Generated Ads
Microsoft has taken a significant step in the world of advertising by creating a minute-long advert for its Surface Pro and Surface Laptop hardware using generative AI. What’s surprising is that the ad was released almost three months ago, and nobody seemed to notice the AI elements. The ad, which went live on YouTube on January 30th, isn’t entirely made up of generated content. Instead, it’s a combination of real footage and AI-generated visuals.
The Creation Process
The team at Microsoft used a combination of written prompts and sample images to get a chatbot to generate text prompts that could be fed into image generators. These images were then iterated on further, edited to correct hallucinations and other errors, and finally fed into video generators like Hailuo or Kling. According to senior design communications manager Jay Tan, "the occasional AI hallucination would rear its head," meaning the creators had to correct some of the AI output and integrate it with real footage.
Choosing What to Generate
When deciding which shots within the ad were to be AI-generated, the team determined that any intricate movement, such as close-ups of hands typing on keyboards, had to be shot live. Shots that were quick cuts or with limited motion, however, were prime for co-creation with generative AI tools. This approach allowed the team to save time and cost while still achieving the desired outcome.
The Role of AI Tools
Microsoft hasn’t specified exactly which shots were generated using AI, but the process involved using AI tools to generate a compelling script, storyboards, and a pitch deck. The team then used a combination of written prompts and sample images to get a chatbot to generate text prompts that could be fed into image generators. Creative director Cisco McCarthy noted that "we probably went through thousands of different prompts, chiseling away at the output little by little until we got what we wanted. There’s never really a one-and-done prompt."
Time and Cost Savings
Despite the iterative process, visual designer Brian Townsend estimates that the team "probably saved 90% of the time and cost it would typically take" to produce the ad. This significant saving is a testament to the potential of AI tools in advertising and content creation.
The Finished Product
The ad has been online for almost three months, and until now, there’s been little sign that anyone noticed the AI output. The video has a little over 40,000 views on YouTube, and none of the top comments speculate that the video was produced using AI. Knowing that AI was involved, it’s easy enough to guess where – shots of meeting notes that clearly weren’t hand-written, a Mason jar that’s suspiciously large, the telling AI sheen to it all – but without knowing to look for it, it’s clear that plenty of viewers couldn’t spot the difference.
Conclusion
The use of AI in advertising is becoming increasingly sophisticated, and Microsoft’s experiment with generative AI is a significant step forward. The fact that the AI-generated elements went unnoticed by viewers until now suggests that, in the right hands, AI tools are now powerful enough to go unnoticed. As the technology continues to evolve, we can expect to see more innovative uses of AI in advertising and content creation.
