Your business & AI: Guidelines for AI in Marketing

Use of AI in marketing is a white-hot topic. Executives, leaders, CMOs, and stakeholders are frantically asking: “How do we use AI to cut costs? Speed up workflows? Generate ads? Supercharge marketing?” Navigating Artificial Intelligence’s role in your marketing is a nuanced topic that we’ve tried to make simple by sharing our distilled, hardline internal AI rules at Frontline.Without making an appeal to the ethical or environmental concerns with AI (which many experts have made), what is the business case for the responsible, long-term position of AI in your marketing? 

It’s critical to first define the functioning of “Marketing” in a business. Marketing isn’t a single activity, or team, or a campaign. The fundamental goal of Marketing is to build trust in your business. The trust built by marketing can take many forms, and be pointed at different targets. Marketing builds consumer trust in your products, services, ability to get results, general dominance over competitors, or signals of overall reliability and longevity. Defined simply: Marketing is: information + proof = trust. Marketing’s trust is also directed at many entities. For example: building trust with customers, vendors, future hires, shareholders, end-users, and investors. 

The problem with AI: AI is dis-trusted by the majority of your consumers. Only 10% of Americans say their excitement for AI outweighs their concerns, subtly implying 100% of consumers have concerns about AI. This puts AI at odds with the key function of marketing, which is building trust.

1. No Fully-AI-generated content. The biggest “business liability” of AI is inaccurate representation with images and text. AI image generation, although improving, will still jumble your text, distort your product, warp your scene, and often represent your brand or product in a damaging way. AI images are commonly associated with shady or fake online scam stores, with surveyed groups reporting up to 46% of responders being burned by an AI product scam. This leads to the following challenges:

• Your customers are associating: “AI images = Scam product.” 
• Your photos may mis-represent the features, functions, or build of your product
• You could be liable for false/inaccurate advertising. 
• Clinical Liabilities: Medical organizations will face intense legal scrutiny for mis-representing their product in a clinical setting.  
• AI “hallucinations” are a well-studied phenomenon. AI has a habit of making up data, surveys, papers, and references in order to achieve your prompt. The liability here should be obvious. 
• 82% of consumers support legislation requiring businesses to disclose AI usage in their marketing.
  
  Note: the difference in using AI-powered tools like “generative image fill” to erase visual clutter, vs. completely re-generating an image to modify the angle, product, or scene.

2. Never use AI to modify someone’s likeness. Distorting a person’s “likeness or appearance” (identifiable, key characteristics) with AI can be damaging for a number of reasons: 

• Legal liabilities around modifying models, stock photos, or using someone’s likeness in an un-authorized way 
• Customers are extremely sensitive to fake profiles. If they notice fake testimonials, user reviews, or photos, it’s automatically perceived as a deception or fraud. 
• Uncanny valley: images of people created with AI will often fall into an “uncanny valley” where customers see them as fake, lifeless, or non-human. Even if they can’t put their finger on why.

3. Only use AI for tasks you could already do manually. Otherwise, AI becomes a crutch, and can quickly build a product you do not understand and cannot edit. 

• If the AI is doing work you literally do not comprehend, your project is 1,000x more likely to run into catastrophic issues with compliance, prompt injection, security concerns, and an inability to edit, correct, or recreate the project.
• Example: if you don’t know how to code, you won’t know how to check your AI-created program for vulnerabilities. If you don’t know how to write, the output of your AI-generated article will be limited by your prompting and editing ability.This guideline keeps AI framed in the helpful, “tool-for-maximizing-speed” realm, and not a dystopian, “I-no-longer-understand-my-own-job” hellscape. 

Frontline has a commitment to accuracy and quality. An AI bot or LLM does not know the goals and nuances of your business in the same way a constant-contact marketing agency does. We’re caretakers of your image, presentation and brand. We are paid to take this seriously. Your products and people in media should be accurate not just for legal reasons but also for public perception. AI generates rapidly, but can be misleading, incorrect, or even make up hallucinations. The legal liability around false information (especially in clinical settings), false product information, and deceptive advertising by use of AI WILL wipe thousands of companies off the map. Don’t be one of them. AI will keep getting better. But your audience is also getting better at noticing it. People are growing tired of AI images and AI-cadence writing (yes, everyone can tell you wrote that article and LinkedIn post with AI). Most importantly? It breaks the social trust we’ve had for thousands of years: People doing business with people. For a number of reasons, your audience doesn’t trust AI. To any marketing director or business leader who insists on using AI to pinch a few pennies, I would love to continue this conversation. Is the adage ‘The customer is always right in matters of taste’ still applicable today? As with anything, there will be a pendulum swing to counter the infinite scrolls of AI-generated content. We know the role of marketing will continue to evolve, and the creative touch of human work will continue to cause companies to stand tall among the chaff. AI is now. Authenticity is the future.  Read more: https://www.pewresearch.org/short-reads/2026/03/12/key-findings-about-how-americans-view-artificial-intelligence/https://theacsi.org/wp-content/uploads/2026/04/AI-Survey-Report_ACSI_2026.pdf