The ADWEEK Guide to AI

AI can perform many of the creative and media functions that power the ad industry, but it isn’t foolproof.

From the reliability of the technology to the quality of its output and its staying power — industry workers don’t fully understand AI’s capabilities, and how it’s going to impact them in the longterm.

And this uncertainty has created a lot of fear around the technology. Here’s why some say that fear is overstated.

While there are vast use cases for AI in the ad industry, there is also a limit to the technology’s capabilities.

One of the most common misconceptions among industry professionals centers around how much AI can be used to replace human ingenuity.

As major brands unleash AI-generated ads and agencies pour hundreds of millions of dollars into AI systems, the first instinct of workers is to think “we’re all going to be out of jobs,” says Ana Milicevic, co-founder and principal at digital advisory firm Sparrow Advisors.

But this is an overstatement, experts believe. AI cannot yet be trusted to run functions autonomously when it regularly hallucinates, and it “doesn’t necessarily understand the nuances of business,” says Domenic Venuto, chief product and data officer of Horizon Media.

He says Horizon Media’s in-house AI platform “enables us to shorten the interrogation of data from days and weeks, into minutes and hours” — but humans are needed to take those insights and translate them for their client’s business needs.

For this reason, Venuto believes that AI doesn’t eliminate the need for media planners and buyers but “changes the nature of their work.”

Likewise for creatives, CJ Bangah, principal at PwC US, says there remains a strong need for human emotion and storytelling, with the “most powerful campaigns” using AI “as the enabler of creatively talented people — versus AI as the lead driver of campaign strategy.”

It does provide opportunities for efficiency gains, and some businesses have reduced their headcount as a direct result of automation technologies. A World Economic Forum survey of more than 1,000 global employers conducted in late 2024 found 41% plan to reduce staff or cut roles that are becoming less relevant in an increasingly automated world.

But the work that AI is taking over is, according to Milicevic, the “kind of work that nobody’s particularly excited about doing.”

Generative AI is especially strong at mining data, summarizing complex information, drafting and translating text and performing calculations.

“While frightening, what this is really going to unlock is different kinds of jobs for all of us,” says Milicevic.

Misconceptions about AI come from a place of fear. Workers argue that AI produces low-quality work that cannot match the creativity of humans as a way to defend their value. Yet the quality of AI’s output is advancing at a rapid clip.

Even now, AI’s output is superior to most humans, believes Ruben Schreurs, group CEO of marketing and media consultancy Ebiquity.

“Is the current landscape of AI or large language models capable of writing a book that’s better than Oscar Wilde or a play better than Shakespeare? Probably not. But can it write better than 95% of the population that isn’t trained or educated specifically for writing or language? I’m sorry, but yes, it can,” he says.

“We don’t like to admit this because we, the human species, like to see ourselves as having something unique,” he adds.

Another misconception that Schreurs is quick to debunk is that AI is a passing trend comparable to how the metaverse was hyped a few years back.

“This is real and it’s ubiquitous, not just in our industry, not just in professional life, but society-wide. There is no denying that this is changing the social construct of work life and private life,” he says.

Automation technologies have been used within the digital advertising ecosystem for decades, powering programmatic advertising and dynamic creative optimization.

Recent advancements in large language models, generative AI and agentic AI have unlocked much greater possibilities in media buying, including granular audience modeling, increased privacy and safety controls and real-time optimization.

Yet handing controls off to AI has come at some cost.

Media budgets are flowing more dynamically than ever before as recent advancements in AI have unlocked the ability for marketers to identify granular audiences and respond in real-time.

Fifty-seven percent of ad buyers surveyed by Accenture towards the end of 2024 said they shift ad spend from one platform to another every 30 days or less — equating to about $35 billion in global ad spend being re-deployed each month.

This may be informed by advancements in media mix modeling (MMM), a statistical analysis method used to determine the impact of various marketing efforts on product sales. Previously, conducting this analysis was a resource-intensive task that most large brands would undergo once a year or so, according to Ana Milicevic, co-founder and principal at Sparrow Advisors.

“We now have the technical capability to make this more of a workflow tool that you revisit monthly and reallocate budgets accordingly, which is not the kind of flexibility that you would have with a traditional MMM,” she says. 

Additionally, automated bidding systems which optimize ad spend according to an advertiser’s goals enable a much more fluid flow of money than manual methods. These have been used for many years to power digital ad spend, forming the bedrock of programmatic advertising.

Recent advancements in large language models (LLMs) and generative AI have enabled the automation of even more of the media buying process. Tools like Google’s Performance Max, Meta’s Advantage+ and Amazon’s Performance+ automate everything from the creative production of ads to their placement and continuous optimization — significantly reducing the time it takes advertisers to launch a campaign.

But the transparency and safety of these tools has come under question, with reports alleging some automated campaign formats have placed ads on harmful websites or caused the inadvertent targeting and tracking of children.

“AI-driven automation can reduce human oversight in decision-making, which can result in ads appearing in inappropriate or brand-damaging contexts. This has been an issue with programmatic advertising, where brands have unknowingly placed ads on harmful or controversial content, leading to reputational risks,” says CJ Bangah, principal at PwC US.

Advertisers sacrifice a level of control and visibility into how their ads are placed when they use AI-powered media buying tools, running the risk of the “enshittification” of the internet, says Milicevic.

“I think the biggest risk is that we remove any scrutiny on which sites we’re buying and end up with machines serving ads to other machines and yet other machines making sure that what’s counted looks like it’s something plausible,” she says.

AI is a powerful tool in the hands of bad actors too, and it has enabled the proliferation of fraud and spam content including made-for-advertising (MFA) websites, responsible for billions in wasted ad spend. According to the Association of National Advertisers’ latest programmatic study, marketers spent 6.2% of their programmatic budgets on MFA sites in 2024 — though this is down from 15% in 2023 as the industry has gotten better at identifying and rooting out spam.

The fix for these new issues is also AI. Scope3, a startup which tracks the carbon emissions of advertising, recently launched a brand safety and suitability product which leverages LLMs to analyze web content against a brand’s specific values, business goals and target demographics — aiming to replace traditional keyword blocklists.

Several household brands have experimented with handing over their creative reins to AI, not just to generate initial scopes but all the way through to the final product.

While generative AI technologies are still in their infancy and have some teething problems to overcome, they are already widely used throughout the creative process and are only expected to grow in influence.

The advent of generative AI technology — including word, image, video and voice generators — is transformative to creative industries, enabling fast and iterative brainstorming and drastically reducing the cost of productions.

Brands including Coca-Cola and Heinz have publicly launched ads created by OpenAI’s text-to-image generator DALL-E, and Ryan Reynolds’ Mint Mobile released an ad script written by chatbot ChatGPT.

AI-generated campaigns are still few and far between right now, and brands who have spent millions of dollars and many years refining their identities are not expected to hand over creative control to AI any time soon.

More often, advertisers and their agencies are using AI to aid with several aspects of the creative process, from generating concepts to testing ad variations for different audiences and geographies.

“It’s not necessarily what you see,” says Christopher Neff, global head of emerging experience and technology at Anomaly. “But it is absolutely transformational in our process. At Anomaly, the vast majority of our people use this technology every day in some way.”

One of the most powerful use cases of AI for creatives is the ability to enable hyper-personalized content at scale — ads which adapt to an individual user based on everything from demographics to location to weather. Rather than manually modifying assets for different audiences, AI can generate dozens of versions in mere seconds from simple prompts. With research showing personalized business interactions drive performance and better customer outcomes, this has proven an attractive business case of AI.

As well as generating content far more efficiently than human labor, AI’s ability to automate time consuming practices such as resizing and adapting assets for different platforms enables marketers and creatives to focus on more strategic tasks.

“AI-powered tools can help streamline the production process…and save marketers’ time, freeing them up from administrative and tactical work to enable them to focus on the higher-level story elements, emotions and customer understanding that will make advertisements more relevant and impactful,” says CJ Bangah, principal at PwC US.

Consumer packaged goods giant Unilever is using more than 500 AI applications across the business, and said in March the technology has driven up to 55% savings and 65% faster turnaround on content, while doubling the click-through rate and holding attention three times longer. 

There are ethical and legal concerns to consider with AI-generated work, which “blurs the lines of ownership and creative authorship,” says Bangah. Dozens of content owners have mounted challenges against AI firms like OpenAI arguing their large language models have been trained on copyrighted material. This explains why there remain limited examples of AI creations in final campaign material.

Another criticism is that AI’s output is derivative and unable to stand up to the quality of human creativity. 

Bangah suggests an overreliance on AI tools in the creative process “could lead to a homogenization of advertising, where brands lose distinctiveness by following algorithm-driven trends rather than bold, human-led innovation.”

Others believe AI can produce exceptional results — provided it is being prompted correctly. 

“If I just put in a prompt and I get a response and I stop there, I believe that is the absolute most detrimental thing to this technological movement, because it is fueling really bad quality and it is creating this lack of effort,” says Neff.

“The thing that I try to teach the most is: how do you push it to take you to places you didn’t know you would go? How do you push it to give you the best possible quality? If creation is no longer as difficult as it was, how do you iterate on it?”