- The new frontier of reputational fraud: deepfake reviews
- How deepfakes work in video reviews
- The rise of video reviews on platforms
- Types of deepfake attacks against business reputation
- How to detect deepfake video reviews
- Legal framework in Spain and the European Union
- Response protocol for a deepfake attack
- Detection and verification tools
- Prevention: shielding your reputation against deepfakes
Your business’s online reputation faces an unprecedented threat: deepfakes applied to video reviews. What just two years ago required advanced technical teams and substantial budgets is now accessible to anyone with access to generative AI video tools. In 2026, the proliferation of video reviews on Google, TripAdvisor and social media has created fertile ground for a new type of reputational fraud combining the credibility of audiovisual content with the sophistication of artificial intelligence.
In this guide, the Maps&Más team analyses in depth how these threats work, what tools exist to detect them, and what response protocol your business should follow if it becomes the target of a deepfake attack. If you have already experienced AI-generated reputation attacks, we recommend reviewing our article on AI and fake review detection on Google.
The new frontier of reputational fraud: deepfake reviews
The online review ecosystem has evolved faster than its protection mechanisms. For years, the battle against fake reviews centred on text: linguistic patterns, publication frequency and suspicious profiles. But the emergence of video reviews on platforms like Google Business Profile, TripAdvisor and Amazon has opened a new attack vector that businesses are only beginning to understand.
Most worryingly, unlike fake text reviews — where detection systems like Google’s already achieve identification rates above 80% — deepfake video detection in reviews is still in an early stage of development on most platforms.
How deepfakes work in video reviews
To protect your business, you first need to understand the mechanics behind this threat. Deepfake video reviews use three main technologies:
Synthetic face generation: AI generative models based on GANs (Generative Adversarial Networks) and diffusion models can create entirely artificial human faces. These faces do not belong to any real person, making legal tracing enormously difficult.
Voice cloning: With barely 10–15 seconds of reference audio, current voice cloning tools can replicate the timbre, intonation and speech patterns of any person. Combined with text generated by language models, the result is an oral testimony indistinguishable from an authentic one.
Lip synchronisation: Current models achieve a precision that deceives the human eye in most cases, especially in low-resolution videos like those typically published on review platforms.
The rise of video reviews on platforms
Major review platforms have invested heavily in the video format in recent years. Google introduced video reviews on Google Maps in 2024 and their algorithmic weight has grown significantly since. A single negative video review — real or fake — can neutralise the effect of dozens of positive text reviews. For businesses dependent on local SEO and Google Maps visibility, the threat is direct and quantifiable.
Types of deepfake attacks against business reputation
We have identified four main categories of deepfake attacks:
- Fabricated negative review: a video with a non-existent “customer” narrating a fictitious negative experience. The most common type.
- Real customer impersonation: the attacker uses a real customer’s image and voice (obtained from social media) to create a video review that person never recorded.
- Coordinated campaign: multiple deepfake videos published from different accounts in a short period to simulate a legitimate reputational crisis.
- Manipulation of real reviews: a positive existing video review is taken and its audio is altered to change the content to negative.
"In the last 12 months we have identified a notable increase in hybrid attacks combining AI-generated text reviews with low-quality deepfake videos. Attackers bet on volume over sophistication, because they know many platforms still lack automated detection systems for manipulated video at scale."
— Analysis Team, Maps&MásHow to detect deepfake video reviews
Although perfect detection does not yet exist, there are indicators that can help you identify potentially falsified video reviews:
Visual indicators
- Irregular blinking: deepfakes often fail to reproduce natural blinking patterns
- Blurry facial edges: pay attention to the transition between face and background, especially around the hairline and ears
- Inconsistent lighting: the light on the face does not match the surrounding environment
- Movement artefacts: sudden movements or momentary distortions when the person turns their head
Content indicators
- Absence of specific details: a deepfake rarely includes verifiable details such as employee names, precise dates, or exact descriptions of the premises
- Generic tone: the narrative is applicable to any business in the sector
- Author profile with no history: the publishing account has little or no prior activity on the platform
Legal framework in Spain and the European Union
The legislation in 2026 offers multiple avenues for action against review deepfakes. For further reading on review manipulation legality, see our guide on whether buying reviews is legal in Spain.
EU Artificial Intelligence Act (AI Act): In force since 2025, it classifies AI systems generating deepfakes as limited risk but with transparency obligations. The use of deepfakes to manipulate consumer opinions may be elevated to unacceptable risk when intent to deceive is demonstrated.
General Data Protection Regulation (GDPR): Creating a deepfake involves processing biometric data (face, voice), which is a special category under the GDPR. Processing without explicit consent constitutes a serious infringement with penalties of up to €20 million or 4% of global turnover.
Unfair Competition Law: Deepfake reviews fit within the provisions on acts of confusion, deception, and exploitation of another’s reputation.
Response protocol for a deepfake attack
If you identify a deepfake video review targeting your business, follow this protocol:
Phase 1 — Documentation (first 2 hours): Download the complete video and publication metadata. Screenshot the author’s profile and history. Record the exact URL and timestamp.
Phase 2 — Technical analysis (hours 2–6): Run the video through deepfake detection tools. Check the coherence of the author’s profile data. Cross-reference video information with your internal customer records.
Phase 3 — Report to the platform (hours 6–12): Use the specific reporting channels for AI-manipulated content (Google, TripAdvisor and Amazon have had specific forms since 2025). Attach technical evidence that the video is a deepfake.
Phase 4 — Legal action (if applicable): File a complaint with the data protection authority for unauthorised processing of biometric data. Pursue unfair competition proceedings if the responsible party is identified.
Detection and verification tools
| Tool | Analysis type | Availability | Estimated efficacy |
|---|---|---|---|
| Microsoft Video Authenticator | Facial manipulation detection | Enterprise & verifiers | High (>85%) |
| Intel FakeCatcher | Facial blood flow analysis | Commercial API | Very high (>90%) |
| Deepware Scanner | Multimodal detection | Free (basic version) | Medium-High (>75%) |
| Sensity AI | Proactive monitoring | Enterprise subscription | High (>85%) |
| Hive Moderation | AI content classification | Commercial API | High (>80%) |
Prevention: shielding your reputation against deepfakes
Real-time monitoring: configure alerts for new video reviews across all your business profiles. Deepfake attacks are often published outside business hours to delay detection.
Verified customer base: maintain an up-to-date record of transactions and customers that allows you to quickly verify whether a “customer” appearing in a video review actually visited your business.
Proactive accumulation of legitimate reviews: a profile with a high and consistent volume of authentic reviews dilutes the impact of any fake review. Consult our guide on how to get Google Maps reviews organically.
Documented crisis protocol: have a specific response plan ready for deepfake attacks, with assigned roles, communication templates and legal contacts ready to act.
Do not wait to become a victim of a deepfake attack. Implement a video review verification protocol today, train your team in detecting manipulated content, and ensure your online reputation strategy includes a specific layer of protection against generative AI threats.
💬 Need to improve your online reputation?
Free consultation with our specialists. Over 2,500 businesses already trust Maps&Más.
Speak to an expert →