Zack Dugow is Founder and CEO of Insticator, a global leader in increasing engagement for publishers.
getty
As artificial intelligence technology becomes more and more powerful, an ever-growing group of industries asks themselves: “How can this technology improve our work?” This is especially true of adtech: Our work has been aided algorithmically for years through options like decision engines and programmatic auctions, so it was a foregone conclusion that our industry would embrace the opportunities provided by enhanced AI.
Those opportunities are plentiful, but they also bring with them challenges and potential pitfalls to watch out for as we leverage this powerful new tool to get the most out of our business. In this post, we’re going to talk about both: the challenges and opportunities that will be coming to our industry, sooner than you think.
Opportunity: AI-Generated Content
One thing that’s been an obvious game-changer for the industry has been the tools to create copy, artwork and even audio-visual content with a simple typed prompt. Instead of delving into artwork libraries in search of the images you need or negotiating licensing for the perfect content, material can be tailor-made to your needs in minutes. In an industry where rapid adjustments often need to be made on the fly to respond to trends, being able to work up an entire campaign’s worth of artwork in minutes is an obvious plus.
Challenge: Losing Your Brand Identity
One issue with AI-generated content, of course, is that it all comes out looking similar. While the technology is becoming more advanced all the time, at the moment the artwork all has a glossy, slightly-plastic sheen that makes it easy to spot, to say nothing of the issues the AI often has with getting human hands right. While less obvious, generative text falls into a similar pitfall, where everybody’s copy ends up coming out with the same middle-of-the-road vibe. In adtech, we stake our claims with bold branding and innovative designs, and so far, AI is better at iterating on what’s already been done. Having good artists and writers on staff can help identify and address the most significant challenges posed by AI.
Opportunity: AI-Aided Decision-Making
One of the great strengths of AI is pattern recognition: taking a big pile of data and identifying trends, categories and predictions. This sort of analysis can take a human operator hours to accomplish, but with an AI in your corner, it can happen in minutes. That ability to process greater volumes of data in shorter amounts of time translates into being able to handle more analysis without having to massively expand your team, providing a significant—and potentially lucrative—business advantage to those companies who deploy it well.
Challenge: Unreliable Data
Last year, U.K. writer Zoah Hedges-Stocks was looking up recipes for fudge and toffee online and stumbled across mentions of a third confection named “scrimgeour.” But after searching for a while, she discovered the truth: There is no such thing. AI-generated recipe blogs had read too much Harry Potter and gotten confused. There’s a character named “Fudge” in the novels, and the AI took the name literally, and it spiraled from there, assuming that “Scrimgeour” was also some kind of candy. The term “garbage in, garbage out” applies here: An AI’s conclusions are only going to be as good as the data that’s fed into it. If you’re not careful about how you collect and vet your data, your conclusions could get very wrong very fast.
Opportunity: Personalization At Scale
One common criticism of generative AI is that it lacks the personal touch. But when you’re running large-scale campaigns, AI can actually provide that personal touch. Crafting a personalized message for each of your customers based on their order history would be both a practical and financial impossibility if you were to have to hire the thousands of copywriters necessary to pull it off. But AI chatbots can perform just that sort of custom, individual work. With outlets like Virgin Cruise Lines offering a personalized, AI-generated invitation from Jennifer Lopez (something the real J. Lo would never have the time to do herself), AI can and will be providing the next generation of “personal touch.”
Challenge: Unrefined Algorithms
"The good thing about computers is that they do what you tell them to do. The bad news is that they do what you tell them to do." - Ted Nelson
We often treat computers and their output as being clinical, unbiased and scientific. But one big flaw to computers is that they’ll do exactly what you tell them to do, not what you meant for them to do. That means that they’ll take their programming in very literal ways, often leading to unintended consequences. Discovering and correcting for these interpretations as they come up is a labor of time, and the best algorithms are going to be developed by houses that can foresee these issues before they happen. Over time, however, more and more of these solutions will be baked into the model.
Final Thoughts
The secret to working with AI is this: Never ask an AI to do something you don’t understand. When employed properly, artificial intelligence can be a powerful labor-saving tool in your kit—but you need to have the ability to check their work. Otherwise, the AI can get things wrong or, worse still, make something up entirely. AI can make it so that you need fewer people on a task to get the job done, but you’ll always need people—and they’ll need to know what to keep an eye out for.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
