The New Era of AI-driven Targeting: Ethical Considerations in a Post-Cookie World

"The times, they are a'changin'," sang Bob Dylan. That's certainly the case in the post-cookie marketing world. Is AI-driven targeting the answer? Let's explore the good, the bad, and the ugly.

Jan 24, 2024

With the recent change in online data privacy standards and the phase-out of third-party cookies, the marketing landscape is going through a major transformation. Marketers, who have long relied on extensive data collection for targeted advertising, face new challenges, and artificial intelligence (AI) is emerging as a potential solution. This article explores how AI might replace cookies in targeting and retargeting and the ethical implications for marketers in this new era.

The TL;DR

While AI can offer significant benefits to marketers in a cookie-less world, it also raises several ethical concerns.

One of the most pressing ethical challenges of AI-driven marketing is its potential to invade user privacy. AI algorithms can collect vast amounts of user data, including their online activity, demographics, and interests. Marketers can use this data to create highly personalized ads tailored to each user.
Another ethical concern related to AI-driven marketing is the issues of transparency and consent. In many cases, users are unaware that their data is being collected and used to target them with ads. This can be problematic, as users may not have consented to have their data used this way.
Finally, AI algorithms have the potential to perpetuate biases in targeting practices. We know that AI algorithms can be trained on biased data, which can lead to biased outcomes. For example, an AI algorithm trained on data biased against women may be more likely to target women with ads for products or services that are stereotypically associated with women.

The End of Cookies and the Dawn of AI Marketing

The third-party cookie has revolutionized digital marketing. However, third-party cookies are increasingly being challenged by those with privacy concerns. In 2018, the European Union introduced the General Data Protection Regulation (GDPR), giving consumers more control over their data use. This has made it more difficult for websites to track user activity across the web and has led to a decline in the use of cookies.

In response to these challenges, marketers increasingly use AI-driven strategies to track user activity without cookies. Artificial Intelligence offers advantages over traditional methods regarding privacy and effectiveness. AI algorithms can process data anonymously by removing personally identifiable information (PII).

They can also identify patterns and relationships that may not be immediately evident to human analysts. Because they can analyze data in real time, they enable quicker and more informed decision-making. In addition, AI can provide personalized experiences to improve user satisfaction and engagement at scale, reducing human error and leading to increased productivity, efficiency, and reduced costs.

The shift from cookie-based to AI-driven targeting is still in its early stages, but it is already significantly impacting digital marketing. By leveraging AI-driven algorithms and machine learning, businesses can effectively identify and engage users who have previously expressed interest in their offerings, increasing the chances of conversion. As AI develops, we expect to see even more innovative and effective ways to use AI for targeting and retargeting.

However, significant challenges also arise, including a lack of data, bias, and complexity. Companies that do not have access to large datasets will also be at a significant disadvantage as these strategies can be difficult to implement and manage. Small marketing teams or singular marketers who don’t have a lot of resources may be unable to keep up.

Despite these challenges, AI-driven digital marketing is a promising trend likely to grow in popularity. Digital marketers who take the time to familiarize themselves with the challenges and opportunities can create more effective and engaging digital marketing campaigns. Let’s dig in.

Leveraging AI for Targeting and Retargeting: The Good, the Bad & the Ugly

As the digital advertising industry evolves to prioritize privacy, AI provides alternative methods for understanding and reaching audiences. Here's how to leverage AI for these purposes, along with the ethical implications of each:

Predictive Analytics

The Good: AI can analyze patterns in first-party data (information directly collected from users) to predict user preferences and behaviors. By examining past interactions with a website, purchases, and engagement with content, AI can forecast future behavior and preferences, enabling more targeted advertising.
The Bad: Using personal data to predict behaviors can raise privacy concerns, especially if users are unaware of how their data is used. Predictive models can limit creativity and diversity in content generation. By relying on past data to predict future preferences, predictive models may reinforce existing trends and patterns rather than exploring new and innovative ideas. This can lead to a lack of originality and homogenization of content, which can be detrimental to the overall quality and richness of the user experience.
The Ugly: Predictive models may not always be accurate, potentially leading to incorrect assumptions about user preferences. There's also a risk of algorithmic bias, where the AI reinforces existing prejudices. For example, a predictive model trained on a dataset that predominantly represents a specific demographic group may generate biased recommendations that favor that group while marginalizing others.

Behavioral Analysis

The Good: Instead of relying on third-party cookies, AI can use behavioral analysis based on first-party data. This approach involves studying user actions on a specific website to understand interests and preferences. For example, if a user frequently views certain types of products on an e-commerce site, AI can infer interest in those products and categories.
The Bad: Ethical use requires explicit user consent for data collection, especially since this involves detailed user behavior monitoring. Obtaining consent can be technically challenging, especially when it comes to tracking users across multiple devices and platforms. This can make it difficult for marketers to ensure they are obtaining consent from all users who are being tracked.
The Ugly: Misinterpreting behavioral data can also lead to incorrect conclusions about user preferences, potentially affecting user experience. When behavioral data is misinterpreted, it can lead to inaccurate assumptions about what users want and need, resulting in products and services that do not meet their needs. This can lead to frustration and dissatisfaction among users, as well as a loss of trust in the organization providing the product or service.

Lookalike Audience Modeling

The Good: AI can identify patterns in existing customer data and use these insights to find new potential customers with similar behaviors and interests. This method allows businesses to expand their reach to new but relevant audiences.
The Bad: This method involves profiling and categorizing users, which can be seen as invasive or a breach of individual privacy. Some individuals may feel uncomfortable with the idea of their data being used in this way, particularly if they are unaware of it or have not consented to it.
The Ugly: There is a risk of excluding or misrepresenting certain groups, especially if the initial user data set is not diverse. For example, if a profiling system is trained primarily on data from a specific demographic group, it may struggle to accurately profile individuals from other groups, leading to potential biases and unfair treatment. There is also great potential for unethical usage, such as targeted advertising, political manipulation, or surveillance. Companies or governments may use this information to influence behavior, shape opinions, or track individuals' movements without their knowledge or consent.

Contextual Targeting

The Good: AI can analyze the content of web pages or videos to place relevant ads. By understanding the context and content consumed by the user, AI can infer the user's interests and serve ads that are more likely to resonate without relying on individual user tracking. Several emerging AI tools use semantic analysis to deliver contextual ads. These tools analyze the content of a webpage or app and then serve ads related to that content. This can be a more effective way to reach users than traditional targeting methods, which rely on user data that may not be accurate or up-to-date.
The Bad: Overemphasis on context can lead to repetitive or irrelevant advertising, potentially impacting the user experience. This can be annoying to consumers and lead to disengagement with a brand.
The Ugly: Avoiding inappropriate or harmful ad placements is imperative for protecting vulnerable populations. AI algorithms often lack important contextual information, such as the user's emotional state, cultural background, or personal beliefs. This can lead to AI-based advertising systems displaying inappropriate or harmful ads to certain vulnerable populations.

AI-powered Semantic Analysis

The Good: This involves understanding the motivation behind user searches, social media interactions, and other forms of content engagement. AI can interpret these data points to determine user intent and interest, enabling more precise ad targeting.
The Bad: Misinterpreting user intent based on semantic analysis can lead to privacy concerns and user discomfort. Contextual variances refer to the specific circumstances and situations in which individuals encounter advertising messages. These variances can include factors such as the time of day, the device used, the user's emotional state, and the surrounding environment.
The Ugly: To avoid inappropriate targeting, the AI must be tuned to understand cultural nuances. If this critical step is neglected, the AI system may inadvertently engage in biased or offensive behavior, undermining its effectiveness and potentially causing harm to individuals or communities.

For instance, certain colors, images, or symbols may hold different meanings or evoke different emotions in different cultures, and the AI must be equipped to recognize and adapt its targeting strategies accordingly.

Dynamic Creative Optimization (DCO)

The Good: AI can automatically tailor advertising creative elements (like images, messaging, and calls to action) in real time to suit individual user preferences and behaviors observed through their interactions on a single site or platform.
The Bad: If marketers fail to ensure that AI-generated content aligns with brand values and messaging, their content may be perceived as inauthentic. AI is a powerful tool, but it should not be used to replace the human touch. Marketers should still be involved in the creative process, ensuring that AI-generated content is relevant, engaging, and on-brand.
The Ugly: Over-personalization can be perceived as invasive and create the impression that a company is tracking and monitoring customers’ every move. Therefore, balancing personalized content with user privacy is crucial. Customers who feel like their privacy is being compromised may lose trust in the company. This can lead to negative perceptions of the brand and its products or services, damage the customer-brand relationship, and make it difficult to build loyalty.

Customer Journey Mapping

The Good: AI can track and analyze a customer's journey from first interaction to purchase. By understanding this journey at an aggregate level, AI can identify key touchpoints for retargeting and effective engagement.
The Bad: Mapping the entire customer journey involves collecting diverse and potentially sensitive data points that raise privacy concerns. Inaccurate or incomplete data can lead to misleading insights and ineffective strategies. Ensuring data accuracy involves implementing robust data collection and validation processes, regularly cleaning and updating data, and addressing any data inconsistencies. This can be technically complex and resource-intensive. It requires the integration of various data sources, the development of sophisticated algorithms, and the deployment of robust infrastructure. Organizations must invest the necessary technology and expertise to implement and maintain these systems successfully.
The Ugly: Ethical considerations also arise when predictive profiling influences decisions at different stages of the journey. Predictive algorithms, trained on historical data, can perpetuate existing biases and lead to unfair outcomes for certain individuals or groups. Factors such as race, gender, socioeconomic status, and other protected characteristics can inadvertently influence the predictions, resulting in unequal treatment and limited opportunities.

To address these ethical concerns, organizations must prioritize fairness, transparency, and accountability in the implementation of AI strategies. This includes regular auditing of algorithms to identify and mitigate bias, providing individuals with clear explanations and control over their data usage, and establishing mechanisms for redress and recourse in cases of unfair treatment.

Available for iOS and Android

Addressing the Ethical Concerns: Self-Regulation is Key

As I’ve discussed throughout this article, the rise of AI presents both challenges and opportunities for marketers. On one hand, AI-driven methods can offer marketers a wealth of new data and insights, which can be used to create more personalized and targeted campaigns. On the other hand, AI raises new ethical concerns, such as the potential for data misuse and privacy breaches.

There are a number of steps that marketers can take to address these ethical concerns. And ensure they are using AI in a responsible and ethical manner. These include:

Enhancing user privacy: Limiting the amount of data collected, using anonymized data, and providing users with more control over how their data is used.
Obtaining user consent: Be sure to obtain user consent before collecting and using their data. This consent should be informed, specific, and freely given.
Increasing transparency: Be transparent about what AI is used to target. This includes disclosing to users how their data is being used and allowing them to opt out of targeted advertising.
Addressing bias: Take steps to address the potential for bias in AI algorithms. This includes using unbiased data to train algorithms and testing algorithms for bias.

Governments around the world are beginning to regulate the use of AI, and every industry, from marketing to healthcare to educational institutions, are rushing to develop standards for ethical AI practices. They are all struggling to keep up with the technology.

As a marketing educator, I know what responsible and ethical use of AI looks like for me, and I believe that self-regulation is key. I was one of the first instructors at my current institution to incorporate an AI Acceptable Use Policy into my syllabus, and other instructors have followed suit. As a marketer, I believe it is up to individual practitioners to take the initiative to do the same.

While we are waiting on legislation and industry standards, marketers can take the following steps to self-regulate AI usage:

Develop a code of ethics: Outline your commitment to ethical AI marketing practices.
Conduct regular audits of AI marketing campaigns: Identify and address potential ethical concerns of any marketing campaigns that use AI for targeting and/or personalization.
Provide feedback to AI vendors: Provide feedback about your experiences with AI marketing tools. This feedback can help vendors to develop more ethical AI algorithms and tools.

As AI technologies continue to develop, it is also important for marketers to stay up-to-date on best practices. By following these best practices, marketers can help ensure that AI is used responsibly to create positive and engaging consumer experiences.

The Bottom Line

Artificial intelligence (AI) in marketing is rapidly expanding, with new applications being constantly developed. AI has the potential to transform marketing in many ways, including:

Personalization: AI can collect and analyze vast amounts of consumer data, which can then be used to create highly personalized marketing campaigns. This can lead to increased engagement and conversion rates.
Automation: AI can automate tasks that human marketers perform, such as data analysis and campaign management. This can free up marketers' time to focus on more strategic tasks.
Predictive analytics: AI can be used to predict consumer behavior, which can then be used to create more effective marketing campaigns. This can lead to improved ROI.

However, there are also challenges associated with the use of AI in marketing, including:

Bias: AI algorithms can be biased, leading to unfair or inaccurate results. This is a particular concern when AI is used for tasks such as predictive analytics or personalization.
Transparency: AI algorithms can be complex and opaque, making it difficult for marketers to understand their inner workings. This can lead to concerns about accountability and responsibility.
Security: AI systems can be vulnerable to cyberattacks, leading to data breaches and other security risks.

It is important for marketers to be aware of the potential challenges of AI in marketing. By understanding these, marketers can use AI to its full potential while mitigating the risks.

In addition to the above, it is also important to consider the ethical implications of using AI in marketing. For example, marketers need to be careful not to use AI to target vulnerable populations or to create misleading or deceptive advertising. Marketers should also be transparent about how they use AI and take steps to protect consumer privacy.

By taking a responsible approach to AI, marketers can help ensure that this powerful technology is used for good.

Like what you just read, leave a comment; if you don’t like it, leave a comment. Either way, leave a comment.