![\"Article](\"https:\/\/dzceab466r34n.cloudfront.net\/Images\/ArticleImages\/136877-OTT-Viewers-ORG.jpg\")
<\/div>\nAI is transforming digital advertising. That is hardly news. But for all of the conversations about AI across the industry over the past 18 months, too many have focused on the positives, and too few have focused on the potential negative impacts.\u00a0<\/span><\/p>\n To be clear, it\u2019s easy to see how AI will be useful for improving campaign performance in digital advertising<\/a>. It\u2019s not easy to look at the potential downsides.<\/span><\/p>\n I\u2019m not just talking about an increase in fraud. The digital ad stack<\/a> can become more efficient while also becoming less transparent, less fair, more fragile, and more centralized, making it less trustworthy and less competitive.<\/span><\/p>\n But beyond business, there are societal risks that are broader and easier to ignore because the costs often land outside the quarterly report.\u00a0 Rising energy use, environmental externalities, and the erosion of consumer trust and authenticity will push the costs of AI outward onto people, culture, and the environment.<\/span><\/p>\n As AI continues to disrupt the industry, it\u2019s critical that we confront these issues with honesty and integrity.<\/span><\/p>\n The environmental cost belongs much closer to the center of this conversation than it usually gets. On the buy side, AI promises ever more automated optimization, ever more creative generation, ever more audience modeling, and ever more real-time decisioning. All of that sounds like software magic. None of it is free.<\/span><\/p>\n The ad industry is talking endlessly about productivity gains and almost not at all about the energy required to produce them. That silence, down from a dull roar twenty-four months ago, is striking. Every additional layer of model inference, automated targeting, synthetic creative production, and optimization infrastructure runs on computing systems that consume power and water. Buyers may enjoy lower labor costs and better campaign performance while remaining almost entirely detached from the environmental footprint required to generate those gains. The benefits are visible in dashboards; the costs are buried in infrastructure.<\/span><\/p>\n This is not an argument against using AI. It is an argument against pretending that intelligence is free.<\/span><\/p>\n For sellers, the same issue appears in a slightly different form. AI-enhanced yield management, forecasting, content classification, moderation, and packaging all increase computational intensity on the supply side. Sellers are told that AI will help them monetize more efficiently, but there is almost no serious conversation about the infrastructure cost of making the whole system more inferential, more automated, and more synthetic.<\/span><\/p>\n Digital advertising has always hidden some of its costs in complexity. AI risks hiding new costs in computation.<\/span><\/p>\n Consumer trust is another casualty in waiting. AI may make creative production easier, but it may also make brands feel cheaper, stranger, or less human. Once consumers begin to assume that advertising is synthetic by default, trust does not merely decline in a single ad; it shrinks in the category.<\/span><\/p>\n That matters because trust is one of the few things in advertising that is both hard to measure and impossible to do without for very long.<\/span><\/p>\n Sellers are exposed to the same trust problem from the other side. If users start to perceive commercial content environments as overly synthetic, overly automated, or full of low-value AI-generated material, then the quality perception of the surrounding media environment can decline as well.<\/span><\/p>\n That is especially dangerous for publishers that rely on a premium context story. AI may help them generate, package, and monetize more content, but if the result feels automated and disposable, the value of the environment itself can erode.<\/span><\/p>\n There is little doubt that AI will <\/span>improve many campaign metrics, lower some labor costs, and speed up many routine decisions. What\u2019s important is that everybody in this industry understands the true costs of those improvements.\u00a0<\/span><\/p>\n Striving for better performance is good. Doing so while sticking our heads in the sand about the societal impact is pure negligence.<\/span><\/p>\n [Editor’s note: This is a contributed article from JWX<\/a>. Streaming Media accepts vendor bylines based solely on their value to our readers.]<\/em><\/span><\/p>\nEnergy and Environmental Cost\u00a0<\/span><\/h2>\n
Consumer Trust and Authenticity\u00a0<\/span><\/h2>\n
Mitigating the risk<\/span><\/h2>\n
\n ![\"Streaming](\"https:\/\/dzceab466r34n.cloudfront.net\/Images\/SiteImages\/457665-2026-Cover-Images-ORG.png\")
\n <\/div>\n<\/p><\/div>\n<\/div>\n![](\"https:\/\/dzceab466r34n.cloudfront.net\/Images\/ArticleImages\/163637-135433-OTT-Quality-Control-1200-ORG-ORG.jpg\"\/)
![](\"https:\/\/dzceab466r34n.cloudfront.net\/Images\/ArticleImages\/163798-monet0-ORG.png\"\/)