The Marketers' Guide to Legal AI in Strategies

At the IAB Public Policy and Legal Summit, panelists shed light on the challenges marketers encounter when incorporating this advanced technology into their strategies.

In the worlds of technology and advertising, Generative AI continues to command attention, as tech giants, agency holding companies, and marketing consultancies roll out new advancements. While it raises questions about future job prospects, marketers perceive it as a solution for burnout and a means to enhance investment in creator content.

Amidst numerous applications and experiments, generative AI confronts a plethora of legal issues and practical challenges. Marketers are tasked with navigating these obstacles as they integrate the technology into their operations, a topic extensively discussed in a panel at the IAB Public Policy and Legal Summit on Tuesday (April 2).

Panelists also highlighted specific definitional distinctions crucial for marketers' understanding, especially as agencies, ad-tech providers, and other platforms rush to embrace generative AI. This clarification is essential as they undergo rebranding and highlight AI functionalities that have been integral to the advertising industry for more than a decade.

Dera Nevin, managing director at FTI Consulting, highlighted, "You've likely employed machine learning and deep learning to segment your audience, develop ad budgets, place ads, and discern which viewers may be more receptive to particular advertisements." Nevin underscored that machine learning and deep learning have long been utilized in the advertising sector and now, the adoption of generative AI for content creation is gaining traction.

Drawing an analogy to cooking, Nevin proposed a metaphor to comprehend AI, wherein algorithms resemble recipes, data inputs act as ingredients, and generated outputs represent prepared dishes. While machine learning equates to a simple recipe, the complexity of deep learning, which propels large-language models and generative AI, mirrors a more intricate culinary process. Just like in the kitchen, the quality of the final product hinges on the quality of the ingredients, and the data on which AI is trained dictates the effectiveness and accuracy of the output.

Nevin emphasized, "To truly grasp the outcome of the interaction between recipes and ingredients, one needs to be aware of what's in the kitchen and who's doing the cooking." However, she noted a lack of transparency regarding the specifics of the recipe and ingredients, asserting that without such knowledge, predicting the resulting "dish" remains uncertain.

Agencies and brands need to be mindful of both the data they input into generative AI systems and the resulting output. When utilizing public-facing generative AI tools like ChatGPT, any data provided becomes part of the algorithm's dataset, whether it's confidential, personal, or otherwise private.

In terms of output, marketers should be cautious of generating acontextual content that may arise when AI lacks contextual understanding. Such output can potentially lead to embarrassing situations for brands. Additionally, as AI aims to mimic human behavior, it may inadvertently click or "behave" in ways that skew metrics or misinterpret engagement levels.

The term "hallucination" has gained traction in describing the unexpected outputs generated by AI. However, Nevin challenges this term, arguing that it humanizes technology inappropriately. These so-called "hallucinations" are the result of underlying mathematical principles and probabilities. While the technology operates as designed, it lacks the human capacity to generate truly original ideas.

Nevin illustrated her point by noting that a human creatively combined two concepts to conceive "Sharknado." She expressed skepticism about AI's ability to accomplish the same feat. However, she acknowledged that AI could plausibly generate sequels such as "Sharknado 2," "3," "4," and "5."

Final Thoughts

The integration of generative AI into marketing operations could gain opportunities and challenges. While the technology offers potential solutions for creative efficiency and content generation, marketers must navigate legal issues, ensure transparency in data usage, and remain vigilant against acontextual outputs. Understanding the limitations of AI, including its inability to generate truly original ideas, is essential. As the industry continues to evolve, careful consideration and strategic implementation will be key to harnessing the full potential of generative AI while mitigating associated risks.