Generative AI has created more opportunities to reduce the cost and time required to produce ads. Yet backlash and legal disputes over AI advertising continue.
Careless use of AI carries a risk of devastating damage to your brand image. Knowing how to use AI's capabilities safely is a new skill that today's marketers need.
This article analyzes real cases of backlash and systematically explains the risks in AI advertising and how to avoid them. It introduces practical ways to maximize AI's benefits while protecting trust in your brand.
Risk 1: the uncanny valley effect and aversion to lifelike AI imagery
Imagery that looks almost human but has unnatural details can trigger the "uncanny valley" effect, a strong sense of unease or aversion. In food advertising in particular, it risks directly damaging the product's image.
A notable past case: McDonald's Japan (August 2024)
An AI ad for french fries drew criticism for unnatural details, including a woman who appeared to have six fingers. It is a representative example of problems that can still occur even as AI generation capabilities improve.
Latest developments and ongoing challenges in 2026
As of 2026, AI-generated anatomical details, such as finger counts and joints, eye focus, tooth alignment, hairlines, and ear shapes, are still not completely natural. In fact, improved AI accuracy is leading to more cases of an even stronger uncanny valley effect: imagery that is almost perfect but subtly unsettling.
Related reading
Online ad types, management methods, and performance improvement: Practical strategies
This article explains in detail how to design a comprehensive online advertising strategy, including AI ads.
The structure of the problem and how it occurs
Topic | Details |
|---|---|
What is the uncanny valley effect? | A psychological phenomenon in which imagery that looks almost human but has unnatural details causes strong feelings of unease or aversion |
Severity in food advertising | Aversion in an ad meant to convey cleanliness and stimulate appetite directly damages the product's image. |
Common problem areas | Finger counts and joints, eye focus, tooth alignment, hairlines, and ear shapes |
How to avoid it: consistently use a hybrid model
Use case | How to use AI | Human role |
|---|---|---|
Brainstorming and rough drafts | Generate large numbers of concepts with AI. | Choose a direction and assess brand fit. |
Visuals without people | Generate landscapes, objects, and abstract imagery. | Check quality and make fine adjustments. |
Final creative featuring people | Do not use AI for the final output. | Have professional creators finish the work manually. |
Risk 2: copyright infringement and why AI training data is your concern
Many generative AI systems use copyrighted images collected without permission as training data. This creates a structural problem: legal risks faced by AI developers directly become brand and legal risks for the advertisers using their tools.
A past case: Asahi Breweries and Stable Diffusion (2023)
Criticism of the brand grew after it emerged that Stable Diffusion, the image generation AI used in its campaign, was the subject of a class-action copyright infringement lawsuit. This issue remains ongoing as of 2026.
Regulatory developments in 2026
The full implementation of the EU AI Act has tightened transparency requirements for AI training data. In Japan, new guidelines from the Agency for Cultural Affairs, released in December 2025, clarify the criteria for assessing similarity between AI-generated material and existing copyrighted works.
Current regulatory status by region (2026)
Region | Legal position | Impact on marketers |
|---|---|---|
Japan | New Agency for Cultural Affairs guidelines (2025) clarify similarity assessment criteria. | Pre-publication checks of generated material are effectively becoming mandatory. |
United States | A Supreme Court precedent established in June 2025 narrows the scope of fair use. | Litigation risk is increasing for commercial use. |
EU | Article 53 of the AI Act explicitly requires training data disclosure. | Training data certificates are effectively essential for ads targeting the EU. |
Criteria for selecting safe AI tools (2026)
Criterion | What to check | Examples of safe tools |
|---|---|---|
Training data transparency | Is there an EU AI Act-compliant training data certificate? | Adobe Firefly, Shutterstock.AI, Getty Images AI |
IP indemnification | Is there indemnification coverage for copyright infringement? | Adobe Creative Cloud Enterprise, Canva Enterprise Plus |
Commercial use license | Does the commercial license state compliance with the AI Act? | Tools whose terms of use explicitly state EU AI Act compliance |
Risk 3: misalignment with brand image and the loss of customer trust through cost cutting
Using AI to cut costs can give the impression of neglecting customers or cutting corners. For brands that emphasize safety, reliability, or a human touch, this risks a devastating misunderstanding that cost savings take priority over customers.
Why judging brand fit is difficult
Since 2024, consumer expectations for transparency and safety in AI use have risen sharply in aviation, finance, and healthcare. Meanwhile, AI adoption is welcomed as a sign of innovation among technology companies and in fashion, creating a growing divide.
A framework for checking brand fit
Check | Question | Decision criterion |
|---|---|---|
Core brand values | What value does your brand promise customers? | Brands that emphasize safety, reliability, or a human touch should be cautious about using AI. |
Fit with the look and feel of AI | Does the look and feel of AI-generated creative match your brand tone? | AI's "mechanical feel" can work against brands that emphasize luxury or warmth. |
The act of using AI itself | Does an "AI-produced" label help or hurt your brand? | It helps tech companies that emphasize innovation, but hurts brands that emphasize tradition or craftsmanship. |
Industry context | How are competitors using AI, and how are consumers responding? | Understand the industry's overall sentiment before deciding. |
Risk 4: lack of transparency and why silence is the worst choice
As consumers become more AI-literate, continued ambiguity about whether AI was used creates unnecessary speculation and is a major factor in escalating backlash.
Trends in consumer AI literacy
Consumers are getting better at identifying AI-generated content, and distrust of companies that conceal AI use is growing. Meanwhile, clear disclosure tends to reduce aversion.
Transparency best practices (2026)
Approach | Example | Expected effect |
|---|---|---|
Disclosure within the ad | "This ad was produced with AI assistance and underwent a final human review." | Prevents unnecessary speculation and demonstrates attention to quality control. |
Sharing the process | "AI generated the ideas, and a professional designer finished the work." | Emphasizes collaboration with creators. |
Establishing a policy | Publish "Our AI Usage Guidelines" on your website. | Makes the organization's responsible approach clear. |
Risk 5: criticism for undervaluing creators and the need for collaboration
As AI advances, organized opposition from the creative industry is growing against efforts to replace creators' work entirely. Companies are being called on to fulfill their social responsibility by establishing an "AI + humans" collaboration model instead of an "AI vs. humans" approach.
The basis of criticism over undervaluing creators
Issue | Specific developments |
|---|---|
Taking jobs away from creators | ArtStation's "No AI" movement and the Japan Creators Association's "FAIR AI" campaign, launched in 2025 |
Exploiting training data without permission | Protests against the use of creators' work for AI training without permission, compensation, or credit |
Disregard for cultural value | Concern about the loss of respect for manual work and craftsmanship |
Building a hybrid model
Stage | AI's role | Human creators' role |
|---|---|---|
Ideas and concepts | Generate large numbers of ideas and explore directions. | Set the strategic direction and assess brand fit. |
Rough drafts and prototypes | Generate multiple options quickly. | Evaluate quality and narrow down the direction. |
Final creative | Serve as a supporting tool, such as for generating backgrounds and assets. | Have professional creators finish the work. |
A hybrid model is both an ethical choice and a practical brand safety strategy. Clearly demonstrating collaboration with creators can significantly reduce the risk of backlash.
Risk 6: platform AI overreach and automated optimization risks in tools such as Advantage+
As of 2026, platform automation features such as Meta Advantage+ and Google Performance Max still risk making unrequested changes that disregard brand intent, violating brand guidelines or delivering unintended messages.
Related reading
AI optimization for Meta Ads: Audience design and campaign strategies that deliver results
This article explains detailed settings for using Meta Ads' AI features safely.
Latest issues with Meta Advantage+ creative (2026)
Feature | Risk |
|---|---|
AI background generation | Automatically generates backgrounds that do not fit brand colors. |
Dynamic product placement | Competitor products may appear in the background. |
Automatic voice narration | Automatically adds a voice that does not match the brand tone. |
Real-time translation | Machine translation that ignores cultural context alters the brand message. |
Checklist for maintaining brand control
Action | Frequency | Purpose |
|---|---|---|
Monitor the ad library. | Daily | Check the quality of automatically generated AI creative. |
Turn off unnecessary enhancement features. | When setting up a campaign | Prevent brand guideline violations. |
Preview every placement. | For each new creative | Check how ads appear on each device and placement. |
Manually review AI-generated assets. | Weekly | Detect unintended changes early. |
Risk 7 (strategic value): AI's real value is producing A/B test variations at scale
Generative AI's greatest value is not making a single ad cheaply. It is generating hundreds of variations in minutes to enable data-driven optimization. This use of AI is a decisive factor in the success or failure of AI ad operations.
The real value beyond cost savings
Traditional A/B testing | AI-assisted A/B testing |
|---|---|
Limited to 2–3 variations | 10 backgrounds x 5 headlines x 3 CTAs = 150 variations generated in a few hours |
Decisions based on intuition and experience | Decisions based on actual user behavior data |
Infrequent testing | Frequent, continuous testing is possible. |
Production costs hinder optimization. | Near-zero production costs enable active experimentation. |
A workflow for using AI's production capacity
Set a hypothesis: Decide which elements to test, such as images, copy, or CTAs.
Generate variations: Use AI to quickly generate large numbers of creative variations and check their compliance with brand guidelines.
Run the test: Monitor the test design during automated delivery optimization.
Analyze the data: Collect performance data and identify insights into why a variation won.
Start the next cycle: Generate new variations based on winning patterns and determine the strategic direction.
By running this cycle weekly, more companies are achieving CPA optimization and increased conversions faster than before.
When using generative AI to produce ads, choosing an environment that lets you generate and edit content in line with your brand's creative rules makes these risks easier to manage. Cascade's Ad Image and Video Generation AI is generative AI built for ad creative. It supports generating banners from product photos and making partial edits through written instructions.
Practical checklist
Preparing to use AI
Brand fit assessment: You have defined your brand's core values and assessed whether AI use fits them.
Tool selection: You have checked the AI tool's training data transparency and IP indemnification coverage.
Internal policy: You have established AI usage guidelines for 2026.
Transparency policy: You have decided how to disclose AI use, such as through a notice within the ad.
During creative production
Hybrid production setup: A human handles the final finishing of creative featuring people.
Quality checks: You focus checks on areas prone to the uncanny valley effect, such as fingers, eyes, and teeth.
Collaboration setup: You have established a hybrid model for collaborating with creators.
Platform operations
Automated optimization monitoring: You regularly check previews of automated optimization features such as Advantage+.
Brand guardrails: You have turned off enhancements that conflict with your brand guidelines.
Testing strategy: You strategically use AI's production capacity for A/B testing.
Risk management
Approval workflow: You have established an internal review and approval workflow for AI ads.
Crisis management: You have prepared a response manual for backlash, with a policy of responding promptly, transparently, and honestly.
Legal measures: You regularly review copyright risks and respond to regulatory changes.
For precautions and steps when creating a landing page with generative AI, see How to create a landing page with AI: Tool comparison and steps.
For an overview of delegating bidding and budget allocation to AI beyond creative production, see What is ad operations automation? Tasks to delegate to AI, platform features, and getting started.
For the difference between producing content with generative AI and delegating ad operations themselves to AI, see What is an AI agent for ad management? Capabilities, differences from traditional tools, and implementation steps.
For ways to help generative AI accurately reference information about your company, see What is LLMO? Differences from SEO, tactics, and measurement.
For an explanation of how ads displayed in ChatGPT work and how ad review is approached, see What are ChatGPT ads (OpenAI Ads)? How to advertise in Japan, pricing, and how they work.
For how to create UGC-style videos with AI and use them in advertising, see What is UGC? Meaning, types, marketing uses, and how to create UGC videos.
FAQ
What is the biggest risk to watch for in AI advertising?
As of 2026, copyright infringement is the most serious risk. Implementation of the EU AI Act has tightened legal requirements, increasing litigation risk for commercial use of generative AI without proof of training data transparency. Choosing a tool with IP indemnification coverage is essential.
Do I need to disclose AI use in an ad?
We strongly recommend it for transparency. As consumers become more AI-literate, concealment creates distrust. A statement such as "Produced with AI assistance and reviewed by a human" can also effectively demonstrate attention to quality control.
Should I avoid all AI ads that feature people?
You do not need to avoid them entirely, but a human creator should always handle the final finishing. Details such as fingers, eyes, and teeth are especially prone to triggering the uncanny valley effect and carry a high risk of damaging brand image.
Should I avoid a platform's automated optimization features?
There is no need to ban them completely, but they may make changes that conflict with your brand guidelines. Maintain brand control by checking previews regularly and turning off unnecessary enhancement features.
Which metrics matter when measuring AI ad performance?
Alongside traditional metrics such as CPA and CVR, brand awareness and sentiment analysis are important. Effective AI ad operations require continuous monitoring of consumer reactions to AI use through methods such as social listening, so you can detect early signs of backlash.
Summary
Success in AI advertising depends not on full automation, but on establishing a hybrid workflow in which humans and AI collaborate. Humans provide strategic hypotheses, ethical guardrails, and the brand's soul. AI provides the scale needed to test those hypotheses. This is the new standard for ad operations in the AI era.
Systematically managing the seven risks (the uncanny valley, copyright, brand misalignment, lack of transparency, undervaluing creators, platform overreach, and misunderstanding strategic value) allows you to maximize AI's benefits while protecting trust in your brand.
It is especially important to use AI as an engine for producing tests at scale, rather than merely as a cost-cutting tool. Generating hundreds of creative variations and enabling data-driven optimization can help you build an advantage over competitors.
For companies considering improving the efficiency of their ad operations, AI adoption is unavoidable. They need to establish appropriate risk management and use AI strategically.
