As many of you reading this now already know, generative AI isn't going anywhere anytime soon. Analysts estimate the Generative AI market will reach $200 billion by 2032, so I'd wager we're still at the very beginning of the evolution of this technology.

Sure, generative AI has been around since the 1960s, but with so many eyes now on the AI prize, AI will likely develop faster over just the next few years than it has in the past six decades.

Knowing that, many across marketing and communications industries have already begun implementing generative AI into their workflows. Around 68% of marketers and 60% of communications professionals use AI at least sometimes in their day-to-day work. Likewise, according to a survey conducted by Billion Dollar Boy, 91% of creators have used generative AI to create content and 92% of marketers have had AI-generated content commissioned.

Most professionals in marketing and communications are already using generative AI, and those who aren't could be at risk of falling behind and losing a crucial competitive edge.

With that as background, OpenAI's launch of a new text-to-video AI tool called Sora is a huge deal. Though the company has said it won't be releasing Sora widely to the public visual artists, designers, and filmmakers who were granted early access to Sora are already showing sneak peeks of what the tech is capable of. Take, for instance, this very convincing elephant made of leaves.

Research from the Conference Board notes that many marketers already use generative AI to summarize content, inspire thinking, personalize content, and conduct research. Soon we will be adding video ideation and creation to the list.

It's imperative to get ahead of the wheel and start thinking now about optimizing text-to-video generative AI to avoid scrambling once everyone's using Sora.

Here's how marketing and communications professionals can begin strategizing for text-to-video AI in efficient, prepared, and innovative ways.

Use text-to-video AI as a starting point

Let's address the elephant—or the five-legged horse—in the room. Text-to-image AI can currently create some startlingly realistic and truly outstanding pictures. However, where a human might know instantly that a horse has four legs, the AI might generate a picture where the beloved equine animal has an extra limb.

Getting anatomy right has been an issue for all kinds of AI-generated images. It seems to be a particular challenge: you've likely already noticed how hands in AI-generated images rarely look like, well, hands.

There are also biases and DEI concerns to consider when sourcing AI-generated images. As any marketer or communications professional can attest, the last thing you want is to unintentionally distribute potentially offensive or harmful content. If you don't specify or acknowledge these biases when using generative AI, it could cause a serious crisis for your company or the brand you're working with.

In addition, it's important to understand how generative AI tools are trained and the intellectual property and moral implications that creates. Many artists have voiced concerns about the theft of their intellectual property, and some have reported seeing AI produce work that appears to have been trained on their art without permission, acknowledgment, or compensation.

Accordingly, using AI-generated content as more than an ideation tool can have reputational repercussions that both you and your clients should be aware of.

At its current stage, AI is best used as an ideation tool—a starting point—not a final product (think previews, not the actual film).

As someone in communications, I know firsthand that it's critical to make sure we're producing well-made content and offerings with as little error as possible. Perfect might be impossible, but careless oversight is equally avoidable. And I'm willing to bet that when text-to-video AI is widely available, it may have similar or different issues to consider.