AI Deepfake Commercial Harm Claims in Europe

1. Meaning and Scope

AI deepfake commercial harm claims concern legal claims arising when artificial intelligence is used to create, manipulate, or distribute realistic but false audio, video, images, or other media for a commercial purpose, causing economic, reputational, privacy, contractual, consumer, or business harm.

Examples include:

fake celebrity endorsements;

fabricated advertisements;

counterfeit executive videos;

fake product demonstrations;

manipulated customer testimonials;

synthetic voices used to authorise transactions;

fake investment endorsements;

impersonation of business owners;

fabricated corporate announcements;

fake news designed to move markets;

manipulated videos damaging a company's reputation;

AI-generated advertisements falsely associating a person with a product;

synthetic images falsely showing a product being used or endorsed.

European law does not currently contain one universal cause of action called a “deepfake commercial harm claim.” Instead, liability can arise under several overlapping legal regimes:

personality and image rights;

privacy law;

GDPR;

defamation/reputation law;

unfair commercial practices;

trademark law;

copyright;

passing off/unfair competition under applicable national law;

contract law;

tort/delict;

product and consumer protection;

platform/intermediary regulation;

the EU AI Act;

collective consumer redress.

2. Why Deepfakes Create a Distinct Commercial Harm

A conventional false advertisement may be recognisably false.

A deepfake can appear authentic.

That changes the legal and economic problem.

For example, an AI-generated video might show a famous actor saying:

“I personally recommend this investment platform.”

If the actor never made that statement, the deepfake could simultaneously create:

consumer deception;

personality-right infringement;

reputational damage;

unfair commercial competition;

financial loss to consumers;

potential trademark or copyright issues.

The same piece of synthetic media can therefore generate multiple causes of action at once.

3. Main Categories of Deepfake Commercial Harm

A. False Celebrity Endorsement

A person's:

face;

voice;

name;

gestures;

appearance;

identity

may be reproduced without authorisation.

The resulting advertisement can falsely imply:

“This person endorses this product.”

Potential claims include:

image/personality rights;

privacy;

misleading advertising;

unfair commercial practices;

passing off or equivalent national unfair-competition doctrines;

damages.

4. Corporate Executive Impersonation

AI can reproduce the voice or appearance of a CEO or company officer.

A fraudster may create:

a fake video call;

synthetic voice instructions;

fake corporate announcements;

fabricated financial statements.

Commercial harm can include:

unauthorised payments;

shareholder losses;

contractual disruption;

reputational damage;

market losses.

Liability may extend beyond the person impersonated to those who knowingly create or distribute the fraudulent material.

5. Deepfake Advertising

A trader may use AI-generated content to falsely represent:

product effectiveness;

clinical results;

celebrity endorsements;

customer satisfaction;

environmental performance;

product safety;

product performance.

Such conduct can potentially constitute an unfair or misleading commercial practice.

The fact that the advertisement is generated by AI does not make the representation less legally attributable to the trader.

6. Fake Consumer Testimonials

Generative AI can produce hundreds of apparently genuine reviews.

Examples:

“I lost 15 kg using this product.”

“This investment platform made me €50,000.”

“My doctor recommended this product.”

If those testimonials are fabricated and presented as genuine consumer experiences, they can potentially violate European consumer-protection principles.

This is particularly serious where the false testimonials affect:

health;

financial decisions;

vulnerable consumers;

children's products.

7. Deepfake Financial Fraud

Deepfakes can be used to impersonate:

bank executives;

investment advisers;

government officials;

corporate directors;

financial influencers.

The resulting commercial harm may involve:

fraudulent transfers;

investment losses;

market manipulation;

unauthorised contracts;

business interruption.

The legal analysis may involve both private-law liability and criminal/regulatory law.

8. Deepfake Corporate Reputation Claims

A fake video may show a company executive:

admitting fraud;

confessing to misconduct;

announcing bankruptcy;

making discriminatory statements;

revealing confidential information.

The company's share price or customer relationships may then suffer.

Potential legal claims include:

defamation;

malicious falsehood;

unfair competition;

economic tort/delict;

market-abuse rules where applicable;

injunctions;

damages.

9. Deepfake Product Demonstrations

AI can create a video showing a product performing better than it actually does.

Examples:

a vehicle apparently completing a safety manoeuvre;

a medical device apparently producing a result;

a cosmetic product apparently producing dramatic effects;

software apparently achieving impossible performance.

If used commercially, such material may constitute misleading marketing.

10. Deepfake and Consumer Protection

European consumer law is particularly important because the consumer may be unable to distinguish:

genuine recommendation

from

AI-generated commercial manipulation.

The legal issue is not necessarily whether the consumer knew the material was AI-generated.

The more important question is:

Was the commercial representation materially misleading or otherwise unlawful?

11. Deepfake and Personality Rights

European legal systems protect individuals against unauthorised commercial exploitation of identity in different ways.

Protection may cover:

name;

portrait;

likeness;

voice;

reputation;

private life;

personal identity.

The exact cause of action differs between Member States.

For example, a person's photograph, voice and public persona may have separate legal protections.

12. Deepfake and GDPR

A deepfake can involve processing of personal data.

Depending on the circumstances, this can include:

photographs;

facial characteristics;

voice recordings;

biometric information;

names;

social-media data.

Where biometric data are processed for uniquely identifying an individual, particularly stringent GDPR provisions may become relevant.

Potential rights include:

access;

rectification;

erasure;

restriction;

objection;

compensation where the requirements for damages are satisfied.

13. Deepfake and Biometric Data

Deepfake creation may require substantial quantities of a person's:

facial images;

voice samples;

videos;

behavioural patterns.

The legal issue is therefore not merely:

“Was the final deepfake harmful?”

It can also be:

“Was the underlying collection and processing of the person's data lawful?”

This creates a potentially independent GDPR claim.

14. Deepfake and the AI Act

The EU AI Act introduces transparency requirements relevant to certain AI-generated or manipulated content.

Deepfake regulation is particularly concerned with making people aware when certain content has been artificially generated or manipulated.

The purpose is to reduce the risk that synthetic media will be mistaken for authentic material.

However, AI Act obligations should not be confused with traditional civil liability.

A violation of an AI Act obligation does not automatically answer every question about:

damages;

causation;

reputation;

contract;

consumer loss.

Those questions may still require other legal causes of action.

15. Deepfake and Defamation

A deepfake stating:

“Company X falsified its accounts”

may constitute a serious reputational attack.

The affected company or person may seek:

removal;

correction;

injunction;

damages;

publication of a correction.

European defamation law differs substantially between Member States.

Therefore, the applicable national law and jurisdiction are particularly important.

16. Deepfake and Unfair Competition

A business may deliberately create a deepfake of a competitor's CEO saying something damaging.

For example:

“Our company is closing because our products are unsafe.”

If the false video causes customers to move to a competitor, the affected business may have claims based on:

unfair competition;

false commercial statements;

interference with business;

defamation;

economic loss.

17. Deepfake and Trademark Law

A deepfake may reproduce:

company logos;

brand names;

packaging;

trademarks.

Trademark law may become relevant where the use:

creates confusion;

damages reputation;

exploits the distinctive character of a mark;

falsely suggests commercial affiliation.

The trademark claim can coexist with personality-right and unfair-competition claims.

18. Deepfake and Copyright

Deepfake creation may also involve copying:

photographs;

films;

recordings;

music;

advertising materials.

Copyright questions can therefore arise at two stages:

Creation

Was protected source material copied to produce the deepfake?

Distribution

Does publication or commercial exploitation of the resulting material infringe protected rights?

Copyright is therefore a potentially important supplementary cause of action.

19. Six Important European Case Laws

Because deepfake technology is relatively new, there is limited European case law dealing directly with AI-generated deepfakes.

Consequently, the following cases establish principles that can be applied to deepfake disputes involving image rights, privacy, reputation, commercial exploitation, data and platform responsibility.

20. Case 1 — von Hannover v Germany

ECtHR, Application Nos. 59320/00 and related jurisprudence

Subject

Publication and use of photographs of Princess Caroline of Monaco.

Principle

Article 8 ECHR protects aspects of private life and personal identity, while the competing freedom of expression under Article 10 must also be considered.

Deepfake relevance

A deepfake can be far more intrusive than an ordinary photograph because it can make a person appear to:

speak;

act;

endorse;

commit misconduct.

Thus, the identity-related interference can be especially serious.

Commercial significance

Where a person's likeness is commercially exploited without consent, the balance between privacy/personality interests and freedom of expression becomes central.

21. Case 2 — Reklos and Davourlis v Greece

ECtHR, 15 January 2009

Subject

Photographs of a newborn child taken without parental consent in a private medical context.

Principle

Control over the use of one's image forms part of private-life protection.

Deepfake relevance

A person's image is not simply a neutral piece of information.

AI can transform an image into an entirely fabricated performance.

Commercial application

If a company takes or obtains a person's photograph and uses it to create an AI advertisement, the image-rights/privacy analysis becomes highly relevant.

22. Case 3 — López Ribalda and Others v Spain

ECtHR Grand Chamber, 17 October 2019

Subject

Workplace video surveillance.

Principle

Technological monitoring can interfere with Article 8 rights and must satisfy appropriate legality and proportionality requirements.

Deepfake relevance

AI deepfake creation often requires extensive surveillance-like collection of images, videos and behavioural information.

The case demonstrates that technological monitoring is not outside human-rights scrutiny.

Commercial significance

Companies using AI to capture, analyse and reproduce employee images must consider privacy and proportionality requirements.

23. Case 4 — Satakunnan Markkinapörssi Oy and Satamedia Oy v Finland

ECtHR Grand Chamber, 27 June 2017

Subject

Publication and processing of extensive personal financial information.

Principle

The Court examined the balance between freedom of expression and privacy/data protection.

Deepfake relevance

AI systems can aggregate enormous quantities of personal information and use it to generate synthetic content.

The case is relevant to the broader question:

When does the commercial processing and publication of personal information become an unjustified interference with privacy?

Commercial significance

Deepfake distributors cannot necessarily rely on the existence of publicly accessible source information as an automatic defence.

24. Case 5 — Google Spain SL v AEPD and Mario Costeja González

Case C-131/12, CJEU, 13 May 2014

Subject

Search-engine processing of personal information.

Principle

Search-engine operators have responsibilities concerning the processing and presentation of personal information.

Deepfake relevance

Modern AI systems similarly:

retrieve information;

aggregate information;

generate content;

rank content;

present information to users.

A deepfake can become especially harmful when algorithmic systems amplify its distribution.

Commercial significance

The case supports the broader proposition that technological intermediaries can have legal responsibilities concerning personal information rather than being automatically treated as neutral conduits.

25. Case 6 — Delfi AS v Estonia

ECtHR Grand Chamber, 16 June 2015

Subject

Liability of an online news portal for unlawful user-generated comments.

Principle

Under particular circumstances, imposing liability on an online platform for unlawful user-generated content can be compatible with freedom of expression.

Deepfake relevance

AI-generated deepfakes may be uploaded or distributed through online platforms.

The case is important when determining whether a platform can simply argue:

“The content was created by someone else.”

Commercial significance

Where platforms exercise substantial control over commercial content, monetisation, moderation or dissemination, their legal responsibilities may become significant.

26. Case 7 — MTE and Index.hu Zrt v Hungary

ECtHR, 2 February 2016

Subject

Liability for user-generated online comments.

Principle

The Court considered the relationship between online platforms, third-party content and freedom of expression.

Deepfake relevance

The same intermediary-liability questions may arise where users distribute:

fake celebrity endorsements;

fake corporate videos;

AI-generated reputational attacks.

Importance

The case illustrates the need to balance:

freedom of expression;

protection of reputation;

intermediary responsibility.

27. Case 8 — Delfi-Related Platform Jurisprudence and Glawischnig-Piesczek v Facebook Ireland

Case C-18/18, CJEU, 3 October 2019

Subject

Injunctions concerning unlawful online content.

Principle

The CJEU accepted that, in appropriate circumstances, courts may impose obligations requiring platforms to address identical or equivalent unlawful content.

Deepfake relevance

This is extremely important for rapidly replicating synthetic media.

A single deepfake may be:

uploaded repeatedly;

edited slightly;

re-uploaded under different accounts;

translated;

algorithmically amplified.

Commercial significance

Effective remedies may therefore need to address not merely one URL or one upload but sufficiently identifiable equivalent versions of unlawful content.

28. Case 9 — Anheuser-Busch Inc. v Portugal

ECtHR Grand Chamber, 11 January 2007

Subject

Commercial intellectual-property interests and property protection.

Deepfake relevance

The case illustrates the broader European recognition of commercially valuable intangible rights.

A deepfake can interfere with:

trademarks;

goodwill;

commercial identity;

intellectual property.

Importance

Commercial identity can have significant economic value deserving legal protection.

29. Case 10 — Ashby Donald and Others v France

ECtHR, 10 January 2013

Subject

Commercial use and dissemination of copyrighted material.

Deepfake relevance

Deepfake creation often depends on existing copyrighted photographs, videos and recordings.

The case illustrates the balancing of commercial expression against intellectual-property protection.

Commercial significance

AI-generated content does not automatically eliminate the rights attached to the source material used to produce it.

30. Consolidated Case-Law Table

31. Who Can Bring a Deepfake Commercial Harm Claim?

Potential claimants include:

Individuals

For:

image misuse;

voice cloning;

reputational damage;

privacy violations;

fraudulent endorsement.

Celebrities/public figures

For:

unauthorised commercial endorsement;

personality-right violations;

reputational damage.

Companies

For:

corporate impersonation;

false statements;

brand damage;

unfair competition;

economic loss.

Consumers

For:

deceptive advertising;

fraudulent recommendations;

financial loss;

misleading product information.

Consumer organisations

For systemic deceptive practices.

Competitors

For:

unfair competition;

false commercial claims;

diversion of customers.

32. Who May Be Liable?

Potential defendants include:

Deepfake creator

Commercial advertiser

AI developer

AI service provider

Person commissioning the deepfake

Advertising agency

Website/platform

Distributor

Marketplace

Trader benefiting from the deception

Liability depends on:

knowledge;

control;

participation;

contractual role;

applicable statutory duties;

foreseeability;

causation.

33. Deepfake Creator vs AI Provider

An important distinction must be made.

Situation A

A person deliberately instructs an AI service:

“Create a fake video of this celebrity promoting my product.”

The person using the system may have obvious responsibility for the resulting commercial deception.

Situation B

An AI provider supplies a general-purpose generation tool that is misused by an unrelated user.

The provider's liability is more complicated.

Relevant questions include:

Did the provider know of the misuse?

Were safeguards available?

Was the system marketed for commercial impersonation?

Did the provider fail to comply with applicable legal duties?

Did it profit from the harmful activity?

Did it respond after receiving notice?

34. Causation in Deepfake Litigation

The claimant must usually connect the deepfake with the commercial harm.

For example:

Deepfake → consumers believe endorsement → purchases increase → celebrity suffers loss of licensing value

or:

Fake CEO video → customers believe company is insolvent → customers leave → company suffers economic loss.

Or:

Fake investment video → consumers invest → consumers lose money.

The more direct the causal chain, the stronger the claim tends to be.

35. Proving That a Deepfake Is Fake

Technical evidence may be crucial.

Evidence can include:

metadata;

source files;

creation history;

AI-generation fingerprints;

audio analysis;

facial inconsistencies;

frame analysis;

blockchain/time-stamping evidence;

platform records;

expert reports;

original recordings.

A claimant should preserve the content before it disappears or is altered.

36. Damages

Potential damages can include:

Economic losses

lost sales;

lost endorsement fees;

lost licensing opportunities;

lost customers;

reduced business value;

fraudulent payments.

Reputational losses

Damage to:

goodwill;

professional reputation;

brand reputation.

Privacy-related harm

Depending on the legal basis:

distress;

loss of control over personal data;

interference with private life.

Corrective remedies

Courts may also potentially order:

removal;

correction;

publication of clarification;

injunction;

cessation of distribution.

37. Injunctions Are Particularly Important

Deepfake litigation differs from ordinary damages litigation because time is critical.

A fake video can go viral within hours.

By the time a final damages judgment is delivered, the commercial harm may already be enormous.

Consequently, interim measures can be extremely important:

emergency removal;

injunction against further publication;

preservation of evidence;

restrictions on further commercial exploitation.

European procedural systems provide different mechanisms for urgent relief.

38. Deepfake and Platform Liability

A platform's responsibility depends upon:

its role;

knowledge;

control;

applicable EU digital-services rules;

nature of the content;

notice;

commercial involvement.

The Digital Services Act has significantly strengthened obligations applicable to online intermediaries and platforms, particularly concerning illegal content and systemic risks.

However:

Platform responsibility is not necessarily automatic liability for every user-generated deepfake.

The precise statutory framework and facts remain critical.

39. Deepfake and Commercial Endorsement

One of the strongest claims can arise when AI creates a false endorsement.

Example:

A famous athlete is digitally recreated saying:

“I use Company X's investment service.”

The athlete never agreed to it.

Potential claims may include:

image rights;

voice/personality rights;

passing off or unfair competition;

misleading commercial practices;

privacy;

GDPR;

trademark;

damages.

Several causes of action can operate simultaneously.

40. Deepfake and Consumer Fraud

Consumers may have claims when the deepfake causes them to enter into transactions.

For example:

AI-generated celebrity video → consumer believes endorsement → purchases financial product → suffers loss.

Potential legal routes may include:

unfair commercial practices;

fraudulent misrepresentation;

contract remedies;

financial-services regulation;

tort/delict;

restitution.

The availability of compensation depends upon the applicable national law and the precise facts.

41. Deepfake and Market Manipulation

A deepfake falsely announcing:

“Company X has discovered a major financial scandal”

could cause investors to sell shares.

Conversely, a deepfake could falsely announce:

“Company X has received regulatory approval for its revolutionary product.”

This could inflate the market value.

Such conduct may engage securities and market-abuse regulation in addition to ordinary civil liability.

42. Deepfake and Freedom of Expression

Not every synthetic video is automatically unlawful.

European law protects:

satire;

parody;

artistic expression;

political speech;

journalism;

public-interest discussion.

Therefore, courts must balance:

Article 8 — privacy/reputation

against

Article 10 — freedom of expression.

The fact that content is false or synthetic is important, but the legal analysis depends on:

purpose;

context;

public interest;

commercial nature;

degree of harm;

identification of the person;

reasonable audience perception.

43. Commercial Deepfakes Receive Less Protection Than Genuine Public-Interest Speech

Where synthetic content is created purely to sell a product or deceive consumers, the creator generally has a weaker claim to expressive protection than someone engaging in legitimate public-interest commentary.

This does not mean commercial speech has no protection.

Rather, the balance may shift significantly where:

the content is deliberately deceptive;

an individual's identity is exploited;

consumers are misled;

serious economic harm is foreseeable.

44. Deepfake and Collective Consumer Litigation

Deepfake scams can affect thousands or millions of consumers.

For example:

An AI-generated celebrity video advertises a fraudulent investment scheme.

If 100,000 consumers invest, individual lawsuits may be impractical.

Collective or representative proceedings can potentially seek:

cessation;

corrective information;

restitution;

compensation;

injunctions.

This makes European collective-redress mechanisms increasingly important in AI-related disputes.

45. Practical Legal Test

A European deepfake commercial-harm claim can be analysed through ten questions:

1. What was fabricated?

Face, voice, video, advertisement, statement, product demonstration?

2. Whose identity or property was used?

Individual, celebrity, company, trademark, copyrighted work?

3. Was consent given?

If not, what legal basis permits the use?

4. Was the material commercial?

Advertising, sales, investment promotion or commercial competition?

5. Was it misleading?

Would a reasonable consumer believe it was genuine?

6. Who created it?

Human, advertiser, agency, AI provider or unknown actor?

7. Who distributed it?

Website, social platform, marketplace or advertiser?

8. What harm occurred?

Economic, reputational, privacy, consumer or business harm?

9. Can causation be demonstrated?

Can the claimant connect the deepfake to the loss?

10. What remedy is most urgent?

Often:

injunction + removal + preservation of evidence + damages

rather than damages alone.

46. Key Defences

A defendant may argue:

“It was satire.”

The court will examine context and whether consumers could reasonably understand it as satire.

“The person is a public figure.”

Public-figure status does not eliminate personality or privacy rights.

“The content was created by a user.”

Platform liability depends on the applicable intermediary framework and the platform's actual role.

“The AI generated it automatically.”

Automation does not automatically eliminate responsibility.

“There was no financial loss.”

Non-economic and personality/privacy harms may still be legally relevant.

“The content was based on publicly available photographs.”

Public availability does not necessarily equal permission for commercial AI exploitation.

47. Strongest Evidence for Claimants

A claimant should ideally preserve:

the original deepfake;

URL and platform information;

date and time;

screenshots;

copies of advertisements;

engagement statistics;

sales data;

customer communications;

original genuine recordings;

licensing agreements;

evidence of lost commercial opportunities;

expert technical analysis.

For a celebrity endorsement claim, a genuine endorsement contract can be particularly powerful evidence of the commercial value of the person's identity.

48. Overall Legal Position

European law is moving toward a multi-layered approach to deepfake commercial harm.

A single AI-generated fake may simultaneously implicate:

**AI regulation

consumer law

privacy

GDPR

personality rights

defamation

intellectual property

unfair competition

platform regulation

contract/tort law.**

The most important principle is:

AI-generated synthetic content does not create a legal vacuum.

If a person or business deliberately uses AI to fabricate an endorsement, deceive consumers, damage a competitor, exploit someone's identity, or cause economic loss, existing European legal doctrines can provide multiple routes to relief.

The principal case-law foundation consists of von Hannover, Reklos and Davourlis, López Ribalda, Satakunnan, Google Spain, Delfi, MTE and Index.hu, and Glawischnig-Piesczek, supplemented by European intellectual-property jurisprudence such as Anheuser-Busch and Ashby Donald.

The emerging legal model is therefore:

Identify the synthetic deception → identify the protected interest → identify the responsible actor → prove commercial or personal harm → establish causation → obtain urgent removal/injunction where necessary → pursue compensation and other appropriate remedies.