AI Marketing


The Definitive Ethical AI Marketing Guide

Oct 25

Julien Palliere

Ethical AI

How do you define AI ethics in marketing? And how do you achieve it? Is it even possible?

Addressing these questions for our marketing clients is one primary function of our work here at Copper Key. These are basic questions we see again and again, and the goal of this guide is to actually provide answers so you can responsibly speak about and use AI in your professional life. In this guide, we will:

  • Define the blockers that we face in marketing adoption
  • Discuss the unblockers we’ve come up with
  • Share Copper Key’s client roadmap to facilitate ethical AI adoption in your marketing team

To dive in, I like the idea of exploring bias as a productive avenue to challenge commonly-held assumptions. (This piece of wisdom comes from feminist anthropology – check out Donna Haraway's Situated Knowledges). It’s no secret: people feel really strongly about AI. In my personal life, I’ve lived through many layman filibusters on the dooms and salvations of AI —these “hot takes” are biases, and none of them have ever helped me succeed in my marketing job. In my professional life, I’ve read many haughty industry reports (Forbes, Microsoft, Hubspot, and Mckinsey). Feel free to skip all of those, because addressing hot takes with B2B speak à la ChatGPT is equally useless. Instead, this guide is informed by 1) our range of AI client consulting projects, and 2) the plethora of academic works that inform our approach (with citations at the end).

So, let’s prod those biases and get at the core of the AI-in-Marketing ethical dilemma. To do this, I’ll set up a definition of AI’s function in marketing:

AI enables the automation of rapid and extensive data collection & analysis for the personalization of a customer’s experience from contact to purchase.

Now I’ll tear it apart, which is easy because that sentence is a legal hotbed. “Data collection,” “data analysis,” “automated personalization,” and “purchase optimization” are terms with very bad ethics PR. That means we as marketers are often afraid of them. That’s okay! This perception shouldn’t deter us simply because it’s normal for evolutions in tech to raise ethical considerations (see why).

Insight #1: The good parts of AI are also the bad parts, and the first step to mitigating risk is explicability.

The Paradox of AI in Marketing: Defining Blockers & Perceived Risk

We’ve all established that AI is a double-edged sword, and that the overwhelming risks cause blockers in our adoption. Fortunately, we’ve fit together perspectives from academic research on AI's risks, what it means to be ethical, and how this translates to your marketing work to address these blockers. 

Firstly, “ethics” is too vague to do any analytical work for us. I like to start with the four cornerstones of ethical ideals that come from real research [see source, Marketing with ChatGPT]:

  • Moral Ideals (Internal Practices): AI’s opacity challenges our ability to scrutinize its results. To counter this, we address the following: Transparency, Accountability, and Responsibility.
  • Legal Ideals (Internal Practices): AI's data reliance turns firms into treasure troves for highly regulated, personally identifiable information (PII). To counter this, we address the following: Sourcing, Security, Storage, and Compliance.
  • Social Ideals (External Practices): AI’s mirroring and amplification of past trends makes it prone to perpetuating dangerous inequalities. To counter this, we address the following: Bias, Participation, Discrimination, and Misidentification.
  • Economic Ideals (External Practices): AI's autonomous capabilities can create job insecurity among employees and cause unease for service users who are unclear or against automated actions. To counter this, we address the following: Delegation, Replacement, and Governance.

Taken together, these ethical ideals serve as starting points for identifying risks and blockers in our marketing work. Just as “ethics” needed to be broken down, so does “AI” (or machine learning, if we’re being technical). This dissection is so crucial and barely happens in marketing literature. If we consult the academic big brains, we find three classes of machine learning: mechanical, thinking, and feeling [source: A Strategic Framework for Artificial Intelligence in Marketing]. Let’s connect our marketing activities with these three machine learning categories:

Breaking down “AI” and “ethics” in marketing helps us expand on the initial definition of AI in-marketing, which is the automation of rapid and extensive data collection & analysis for the personalization of a customer’s experience from contact to purchase. Just as I pointed out the legal headaches from that initial definition, we can now commit to scrutinizing the specific practices we outlined. Here are the ugly truths:

  1. Negative Impacts on Business (Internal Blockers). This perspective places emphasis on internal or operational liabilities that can arise from the use of AI. Examples include data breaches, algorithmic bias, agency costs, and regional data privacy regulations. These risks can negatively impact a firm’s financial situation, regulatory compliance, workplace culture, and more..
  2. Negative Impacts on Consumers (External Blockers). This perspective places emphasis on the consumer experience. It’s important to mitigate these issues, as public sentiment can quickly translate to cancel culture or shareholder pressures. Examples of negative consumer impacts include privacy concerns, lack of AI transparency, and discrimination. These risks can negatively impact a firm’s customer loyalty, brand image, reputation, and more.

Alright! Now that’s a series of issues we can work with. Breaking down these issues, consequences, and their related ethical considerations into our industry’s vocabulary makes tackling these issues easier — and more effective. 

Insight #2: You manage your job through a series of concrete processes, and so does AI. Break down the insurmountable obstacles by naming and explaining exactly what you do — and how AI fits in.

Unblocking AI in Marketing: Mitigating Risk through Explicability

Upon closer look, none of the problems outlined above are completely new to us marketers. In fact, we’ve dealt with much of this before. Personally, establishing some no-brainer correlations has helped me feel grounded as we explore these issues in our consulting projects. Some of these correlations include:

  • Security & Compliance: The dot-com boom of the late '90s made cybersecurity a widespread priority, and the global costs of cybercrime are projected to reach $10.5 trillion by 2025. In response, internet companies implemented robust security protocols and regular audits. Source: Embroker
  • Discrimination & Participation: Social media platforms betrayed the dangers from machine learning’s widespread reification of existing societal biases. In response to criticisms, some social media platforms adopted third-party audits for their content moderation practices. Source: Forbes
  • Transparency & Accountability: Mobile technology was notoriously spotlighted for rampant consent and privacy breaches. One result is that new industry regulation required marketers to obtain customer consent for data handling and mobile marketing messaging. Source: ANA

Insight #3: Marketers understand that “newness” is just a buzzword, and that precedent can always be found underneath. The new problems that AI poses have old solutions.

Again, academic literature already has suggestions for adapting this precedent into organizational learnings and marketing design practices. From consumer sentiment analyses and external auditing to external AI ethics resources and team workshops, there are many processes I’ve employed both for clients and within our own marketing firm.

Adopting AI in Marketing: The Copper Key Roadmap

The plethora of mitigating strategies responding to each of these developments extends to AI. As a sub-brand of the award-winning marketing agency, Column Five, we have more than 15 years of experience in navigating the developments and pitfalls of marketing. I’ll get straight to it – this is the roadmap we use with our clients on AI implementation projects:

Interested in putting this into practice in your company? We’ve created a product package that does just this: Ethics Charter

There we have it! From AI sensationalism to real steps you can take in your team’s path to AI adoption, we’ve unpacked a lot of important facts about AI ethics in marketing. I’d love to hear from you on whether this relates, what ideas this prompted, and what’s missing. In the meantime, keep an eye on upcoming blogs where I’ll be using our project experiences to provide concrete context for what we’ve talked about here. 

Happy generating,

Helpful resources

Ethical AI