We are witnessing AI reshaping many industries in real time, and the advertising industry is no exception. With the ability to reduce cost, streamline processes, and improve ad performance in totally new ways, we may see more and more money flow into the advertising ecosystem based on better performance and attribution. We should be seeing a rising tide lifts all boats scenario, but that’s not what’s happening.
As advertising buyers increasingly embrace AI to optimize campaigns, discover audiences, and improve attribution, a crucial question emerges for publishers: are they seeing a benefit? The ad industry’s “buy side” is rapidly leveraging AI to drive efficiency and performance, yet the “sell side” is simultaneously seeing lower traffic volume due to consumer adoption of AI platforms. Also, while AI promises to automate tasks and drive unprecedented efficiency, the initial and ongoing expense of building and maintaining these systems may create a paradox where the tools meant to save money become a major expenditure themselves. Numerous public studies have recently questioned the ROI of these projects so far. The third paradox touches on sustainability: is this powerful technology a solution or an accelerant to our environmental problems? To truly move forward as an industry, steps must be taken to confront and find solutions for each of these paradoxes.
Publishers versus media buyers
The most urgent of the three paradoxes to address is creating a balanced ecosystem. Publishers are the bedrock of the ecosystem; if they collapse, there will be nothing left to maintain. As AI helps media buyers optimize ad performance by reducing waste and targeting the correct consumer, it should drive more money into the ecosystem and, in theory, to publishers. But, with many users just searching for answers in the AI summaries on web pages or in AI platforms, traffic is being cannibalized from publishers. So far, it’s not clear that the decline in traffic is also driving a revenue problem (those advertiser budgets still get spent on publishers). However, as LLMs launch ad platforms, those products may divert advertiser dollars from publishers to LLMs. Perhaps growing advertiser budgets will leave enough for both parties, but those dynamics aren’t clear.
However, LLMs depend on publishers to feed their content. Instead of treating AI as an adversary, some publishers and AI platforms are finding a path toward a symbiotic relationship. As AI models reference and link back to publishers’ trusted sources, they drive users to the original content, creating a new and potentially more sustainable traffic stream. This approach acknowledges the value of the high-quality content that feeds the large language models and provides a form of fair compensation and attribution. This shift is gaining traction as industry bodies, such as the IAB Tech Lab, step in to create frameworks like their new AI Content Monetization Protocols (CoMP). These protocols aim to protect intellectual property and create a structured ecosystem for monetization, ensuring that both publishers and users are part of a sustainable and transparent future.
Cost-effectiveness versus implementation
Whenever there is a new technology being deployed across organizations and at the scale AI is being deployed, there is an expectation to see an ROI on that technology. However, massive organizational changes take time to implement, and AI is no different. For all the benefits AI offers, be prepared for upfront costs in spending, time, and resources—not only to develop your own technology but also to train employees on the new tools, just as you would with any other massive technological shift. Businesses should not be treating AI any differently than other business investments. Nearly eight in ten companies reported using generative A.I. but just as many have reported ‘no significant bottom-line impact’.
If you were investing in a new warehouse for storage to increase your supply to keep up with demand, launch a new product line, and have capacity to keep it in stock, all of these things would have clear and measurable objectives. Everything about the new warehouse would ladder up to a long-term goal to drive growth and success for your business. You also wouldn’t expect your new warehouse to be up and running in a week or two, you would have a clear timeline and benchmarks to hit. The same focus, attention to detail and measurable goals need to be implemented with AI in your business.
There seems to be an urgency to have AI embedded all across your organization, but you wouldn’t do that with other major organizational changes, as you wouldn’t want to rush into things. The same care and considerations need to be taken with AI. This can be done while simultaneously funding small experimental investments to learn more about where the technology might help your organization.
Sustainability versus increased energy use
With all new technology comes the conversation on sustainability. AI is no different in that matter, and there are noted sustainability concerns. In fact, The International Energy Agency (IEA) projects that global electricity demand from data centers, heavily driven by AI, will more than double by 2030 to over 945 TWh, roughly equivalent to Japan’s total electricity consumption today. In the United States, data centers are on course to account for almost half of the growth in electricity demand. It’s not clear there will be a real solution to this, but we have two trajectories of hope. One is the simple continuation of the shift of our electricity grid to more sustainable sources, which is moving forward at varying paces around the world. The second, more moonshot hope, is that AI itself will help us discover new and sustainable options, such as helping design more efficient batteries and solar panels.
Despite a number of paradoxes in the industry right now, there are clear steps that can be taken. Right now we are in the phase of people pointing out the problems with this new technology, but the time for that is over. AI has the potential to become a net positive across all of these paradoxes as well as ones we may not even be aware of yet, and it is time to shift to a new way of thinking with more problem-solving ideals.
