Abstract
- FTC focus. The FTC has begun to outline what AI abuse looks like and is already working on e-commerce cases involving AI abuse.
- Risk complexity. The combination of the ephemeral nature of service provision and the scale of AI complicates identifying fraud risk for marketers.
- Ethical inquiry. Marketers must consider the ethical implications of their AI strategies to ensure customer trust and compliance.
It's no exaggeration to say that the adoption of AI-based products and services is accelerating rapidly. The popularization of tools such as ChatGPT, Propensity, and Bard (now Gemini) has created immense curiosity to learn how to use just about anything related to or with AI, and important information on AI ethics. Considerations arise.
AI ethics: Balancing curiosity and truth
Marketers of products that include AI are looking to capitalize on that curiosity. But when does courting a customer cross the line into false advertising? What ethical concerns weigh the risks and how a customer experiences her AI-based product or service? Will it happen?
With AI, marketers need to more directly identify and understand the benefits of AI-based products. Campaign tactics that don't clearly explain the benefits of your product or service can overpromise and mislead customers with unmet expectations.
Related article: AI, privacy, and law: Unraveling the US legal framework
FTC warning
Like many organizational and industry leaders, the FTC has been closely following developments in AI, but has also been concerned about market transparency. Last year, Michael Attleson, an attorney in the Federal Trade Commission's (FTC) advertising practice division, said that AI-based products are overhyped and that some caution is needed to balance the concerns. posted an FTC notice online expressing concerns.
FTC's 4 Key Questions About AI Ethics
This letter outlines four key questions the FTC uses to consider the effectiveness of AI-based solutions.
-
Are you overstating what your AI products can do?
-
Do you promise that your AI product will perform better than non-AI products?
-
Are you aware of the risks?
-
Does the product actually use AI?
Related article: Will the AI Executive Order be the data privacy compass for customer experience?
The value of AI: Proof of enhancement and impact
All are great questions. I imagine this last question will come up frequently among marketers of AI-influenced solutions, as many popular software solutions are incorporating AI tools at a furious rate.
But proving the value of AI-enhanced products will also be particularly difficult. How do consumers know that an enhancement has significantly improved their experience with a product or service?
Related article: FTC won't tolerate generative AI deception in marketing and customer service
FTC cracks down on AI misrepresentations
Cases have already demonstrated the significant risks of AI-based experiences not delivering as promised. In February, three business coaches settled FTC allegations that they deceived wealthy e-commerce clients with unsubstantiated promises of increased consulting revenue. Part of their services included operating online sites on behalf of clients that were featured on networks such as Walmart and Amazon. The coaching was touted to include AI-powered services for the field. In the end, the vast majority of customers did not achieve the promised returns, and Walmart and Amazon suspended many sites for policy violations. Under the FTC settlement, the accused coaches had to return nearly $21 million in assets and accept a permanent ban from consulting in the e-commerce field.
Related Article: Executive Order on AI: Necessary Action or AI Governance?
AI Ambiguity: Navigating Benefits and Claims
As explained in Attleson's letter, AI encompasses a variety of frameworks, making the definition of benefit very vague. Bad actors often exploit this ambiguity to sell ineffective products to unsuspecting customers. Consumers must be able to measure or compare benefit claims. For example, a health drink may claim to have iron as an ingredient, but it may not contain enough iron to actually benefit your body. Consumers can compare the amount of iron in one drink to another by comparing nutrition labels.
However, many products are services, and the customer experience with services is temporary. It can be more difficult to determine whether such a temporary experience will yield results.
Related article: AI in customer experience: Impact on the customer journey
5 AI ethics questions marketers should ask about their strategies
So what should marketers consider when implementing AI in an ethical way to ensure customers see the benefits rather than the drawbacks, as in the FTC case?
Answering the following questions will reveal where marketers should focus their efforts to maintain an ethical customer experience.
AI Ethics Question 1: What social aspects of AI functionality should be the responsibility of algorithms?
The answer to this question addresses decision-making when human intervention is essential. AI capabilities should not be solely responsible for making decisions that have a significant impact on society. AI algorithms are trained on large datasets that can reflect social biases. The ability to quickly process many decisions based on biased data can amplify discriminatory or unfair outcomes.
Human resources decisions, such as hiring or firing employees, or services that require customer approval, such as consumer bank loans, are prime examples where automated AI-based processes can benefit from human intervention.
Related article: AI in Marketing: Balancing creativity and algorithms for marketers
AI Ethics Question 2: What personal information should AI have access to to determine outcomes?
The answer must relate to how personally identifiable data is handled within the marketer's organization. The amount of personal information accessed by an AI algorithm should be limited to the amount necessary to achieve its intended purpose. For example, if AI recommends a product to a customer, it may need access to the customer's purchase history, but not the customer's social security number.
Effective data privacy depends on permission. Organizations must ensure that AI models consistently process only customer data that they have permission to access.
Related article: 2024 AI Roadmap for Marketers
AI Ethics Question 3: How can we ensure that AI is not trained on biased data and perpetuates discriminatory practices?
These kinds of questions are at the heart of many debates about the use of AI that have made headlines over the years, such as applications in facial recognition. Concerns about AI decisions based on biased data are also why developers are working on modeling techniques to reduce bias.
One example I profiled is Latimer, a large-scale language model for research. It is designed to include cultural data in AI models to teach them how to eliminate cultural discrimination in their models.
Marketers need to stay current with the latest advances in AI, including research on the optimal deployment of search augmented generation (RAG), to understand practical options for ensuring fair training in AI applications. there is.
AI Ethics Question 4: How will customers understand how AI-powered recommendations and decisions are made?
In contrast to the FTC e-commerce case that over-promised and under-delivered, how do customers interact with AI-based outcomes to ensure that the AI-enabled customer experience feels authentic and reasonable? It is essential to evaluate the
There are a number of simple actions marketers can take along the buying process to highlight customers' options for managing their interface with AI. For example, you need to make the customer's options prominent so that they know where they can opt out of interacting with her AI chatbot if they wish. Options should be transparent so customers can understand them well.
AI Ethics Question No. 5: Are there any potential social impacts from the impact of AI in our products or services?
Taking the time to brainstorm potential social impacts should be an essential part of developing a product or service when AI is involved. For example, introducing AI into many processes could lead to the automation of many jobs, including the elimination of some positions. A key question to ask is whether the people affected live in areas where the brand has already made significant investments. Brands will likely counter previous positive efforts to shore up financial support with negative, attention-grabbing messages. It is important to highlight potential impacts in your mind map and determine steps to mitigate them.
final thoughts
AI-infused marketing technology is rapidly becoming a key element in meeting customer expectations in the marketplace. Marketers are now responsible for ensuring that these expectations are met to the highest standards.
