Market impact of the EU AI Act’s transparency obligations: vertical benefits and competitive dynamics

Introduction: The EU AI Act and transparency obligations

The EU’s Artificial Intelligence Act (AI Act) entered into force on 1 August 2024 and introduces a risk‐based regulatory framework for AI systemsdigital-strategy.ec.europa.eu. Article 50 – the Act’s transparency chapter – imposes additional obligations on providers and deployers of certain AI systems. From 2 August 2026, providers must tell users when they interact with an AI system, mark AI‑generated outputs in a machine‑readable way so that other systems can detect them, and label deepfakes and AI‑generated text used to inform the publicartificialintelligenceact.eu. Business using AI chatbots, AI image/video generators, or emotion recognition systems must notify users and ensure that the outputs carry visible and machine‑readable labelsweventure.de. The Act distinguishes four risk categories—prohibited, high‑risk, limited‑risk and minimal‑risk—with different obligationsnoticetheelephant.com. Generative‑AI systems used in marketing, customer support and e‑commerce are usually limited‑risk, meaning transparency obligations apply but there is no blanket bannoticetheelephant.com. High‑risk applications (e.g., AI credit assessments or recruitment tools) require strict documentation, risk management and human oversightgoodwinlaw.com. Prohibited practices include manipulative systems, exploiting vulnerabilities of children, and social scoringnoticetheelephant.com.

The EU Commission is drafting guidelines and a code of practice to standardise how AI‑generated and manipulated content is marked (e.g., through watermarks or metadata), and the Commission invites stakeholders to contributedigital-strategy.ec.europa.eu. Deepfake labelling requirements apply earlier (from 2 August 2025)artificialintelligenceact.eu.

Summary of risk categories and key transparency duties

Risk category	Examples of AI systems	Key duties under Article 50 and related provisions	Penalties
Prohibited practices	AI systems that manipulate people’s behaviour, exploit vulnerabilities of children/persons with disabilities, or enable social scoringnoticetheelephant.com	Total ban; no deployment	Up to €35 million or 7 % of worldwide turnover for violationsnoticetheelephant.com
High‑risk	AI used in critical areas: credit & insurance, employment, education, law enforcement, border control, medical devices, and voting**goodwinlaw.com**	Providers must establish risk & quality management systems, use high‑quality datasets, document design decisions, ensure human oversight and pass conformity assessments before placing on the marketgoodwinlaw.com; deployers must inform individuals when using emotion recognition or biometric categorisation systems	Up to €35 million or 7 % of worldwide turnovernoticetheelephant.com
Limited‑risk	Most generative‑AI systems used for marketing, chatbots, e‑commerce recommendation engines, corporate communications tools, language modelsnoticetheelephant.comamalytix.com	Providers and deployers must inform users they are interacting with AI (e.g., chatbots), label AI‑generated content with visible and machine‑readable markers, and notify people when emotion‑recognition or biometric categorisation is usedartificialintelligenceact.eu	Proportionate penalties; non‑compliance may lead to fines up to €7.5 million or 1.5 % of turnover for SMEs, or higher for large companiesartificialintelligenceact.eu
Minimal‑risk	AI used for spam filtering, simple analytics, productivity tools	Subject only to voluntary codes of conduct; no specific transparency requirements.	n/a

The remainder of this report examines how transparency rules will affect specific vertical markets, the benefits that may arise, and how competition dynamics may shift across industries.

Figure 1 – graphic overview of risk categories and their application across selected verticals. Most generative AI applications in marketing, social media and e‑commerce fall in the limited‑risk category, while high‑risk systems are concentrated in finance, healthcare and employment.

Market impacts and vertical benefits

1 Marketing & advertising

Obligations and operational adjustments

Marketing tools such as chatbots, content generators, lead scoring models and programmatic advertising platforms fall within the limited‑risk categorynoticetheelephant.com. From 2 August 2026, marketers must:

• Notify users when they interact with an AI system (e.g., disclose that a chatbot or virtual assistant is AI‐driven)artificialintelligenceact.eu.
• Label AI‑generated copy, images, videos or voices, both visibly and through machine‑readable metadataartificialintelligenceact.eu.
• Clearly mark deepfake advertisements or AI‑generated political messagingrealitydefender.com.

Existing marketing practice already emphasises human review and editing. Quibble’s marketing article notes that from August 2025 onwards, unedited AI copy could lead to compliance issues and poor performance; marketers should keep human editors “in the loop”, audit AI tools and transparently disclose AI usagequibble.digitalquibble.digital. Surveys by IAB found that over 60 % of marketers support labelling AI‑generated ads; concerns include misinformation and loss of creative controliab.com. Many marketers have experienced AI‑related incidents such as hallucinations and bias, which led to paused campaignsiab.com.

Benefits and competitive implications

Restoring consumer trust: Transparent labelling can help rebuild trust in digital advertising, where deepfakes and targeted disinformation threaten brand safety. Consumer research suggests that people respond positively to transparency: a global survey on AI‑generated video content showed that 20.6 % of consumers would engage more with brands if they disclose AI usage, and 85.5 % prefer when AI‑generated videos are assisted by humansheygen.com. Another article notes that clear disclosure, human oversight, and metadata watermarks help build trustecija.com. Thus, transparent marketing can improve engagement and reduce legal risk.

Differentiation and responsible branding: Complying with the AI Act allows companies to position themselves as ethical and responsible, differentiating from competitors who rely on opaque AI. TrustPath emphasises that transparency fosters accountability and bias mitigation, offering reputational benefits and improved model developmenttrustpath.ai. Carbon6 notes that clear disclosure in e‑commerce not only meets legal requirements but can improve customer trust and create competitive advantagecarbon6.iocarbon6.io. Marketers who invest early in compliance can therefore gain first‐mover advantage.

New markets for detection and watermarking services: The Act catalyses demand for tools that add or verify digital watermarks and content credentials. The policy paper on multimedia authenticity lists emerging providers such as Friend MTS and Google SynthID that watermark synthetic media to identify manipulated contents41721.pcdn.co. Advertising agencies may partner with these vendors to ensure compliance and to reassure clients that content is authentic.

Competition and burdens on SMEs: While large advertising platforms have resources to implement labelling standards, smaller agencies may struggle. The Digital SME Alliance warns that the lack of technical standards and regulatory sandboxes could leave SMEs facing high compliance costs and legal ambiguity, potentially distorting competition in favour of large incumbentsdigitalsme.eu. However, the AI Act introduces proportionate fines and reduced conformity assessment fees for SMEsartificialintelligenceact.eu, and it ensures their participation in standard settingartificialintelligenceact.eu, mitigating some competitive disadvantages. It also encourages digital marketing service providers to offer AI governance and compliance support, creating new B2B opportunitieseverestgrp.com.

2 E‑commerce and retail

E‑commerce platforms and sellers using AI for product recommendations, chatbots, content creation or visual try‑ons are also limited‑risk systems. Articles aimed at Amazon sellers explain that from 2 August 2026 they must attach machine‑readable metadata and visible labels to AI‑generated product descriptions, images and videosamalytix.com. Sellers must also inform customers when they interact with AI in chatbots, and label any AI‑generated or manipulated reviews or testimonialsamalytix.com. Ecija’s legal analysis lists retail and e‑commerce among the sectors most exposed, warning that insufficient labelling or superficial warnings may lead to penaltiesecija.com.

Benefits and competitive implications

Enhanced consumer trust and conversion rates: Transparent labelling may initially slow marketing processes but can improve long‑term conversion rates by signalling authenticity. Carbon6 argues that clear disclosure fosters trust and “ethical AI use” is a way to differentiate and retain customerscarbon6.iocarbon6.io. This is consistent with HeyGen’s findings that a significant share of consumers (nearly 70 %) are comfortable with AI‑generated videos but value transparencyheygen.com.

Operational adjustments and compliance costs: Sellers must integrate metadata fields and update product information management systems to flag AI‑generated content. They may need to invest in watermarking tools or partner with providers. Start‑ups and small sellers may face disproportionate costs; however, the Act’s proportional fee structure and SME‑friendly fines offer some reliefartificialintelligenceact.eu.

Fairer marketplace competition: Transparent labelling discourages the use of deceptive AI‑generated reviews or manipulated images, levelling the playing field between honest sellers and those using generative AI to mislead. It also mitigates the risk of reputational damage from unlabelled deepfakes or hallucinations. For platforms, implementing detection and removal mechanisms can reduce liability and maintain consumer confidence.

3 Media, entertainment and corporate communications

Media organisations, streaming platforms and gaming studios often use generative AI to produce news articles, dubbing, deepfakes, or synthetic actors. From 2 August 2026, they must clearly label AI‑generated content and watermark synthetic media. Deepfake labelling applies earlier (2 August 2025). RealityDefender notes that labelling obligations extend to all industries distributing AI content—media, advertising, gaming, telecommunications, corporate communications, education, etc.—and apply to both developers and deployersrealitydefender.comrealitydefender.com. Ecija emphasises the need for visible warnings and metadata marking and warns that superficial labels or lack of oversight could lead to sanctionsecija.com.

Benefits and competitive implications

Combating misinformation and restoring credibility: Generative AI blurs the distinction between synthetic and real content, causing a “credibility crisis” for governments, businesses and journalistss41721.pcdn.co. Transparency measures, coupled with verification tools and digital literacy efforts, can help rebuild trust in media. By clearly marking AI‑generated news or entertainment, broadcasters can protect viewers and avoid reputational damage.

Innovation in content verification and watermarking: The rise of deepfakes spurs demand for verification technologies. The policy paper lists companies developing watermarking and content authenticity solutionss41721.pcdn.co. Adoption of such tools will become a competitive differentiator for streaming services and news outlets seeking to reassure advertisers and regulators.

Diverse competition outcomes: Major media houses might absorb compliance costs easily, while independent newsrooms and smaller content creators may struggle. However, transparent labelling could level the playing field by exposing AI‑generated clickbait and reducing the advantage of unscrupulous actors. Early compliance may attract quality‑seeking advertisers and platforms may compete on being “trustworthy” and “human‑driven”.

4 Social media and technology platforms

Large social media platforms host vast amounts of user‑generated content, including AI‑generated posts, images and videos. As deployers of AI, they must enable detection, labelling and removal of unmarked deepfakes and manipulated media. The AI Act’s territorial scope means platforms outside the EU must comply if their services reach EU usersrealitydefender.com.

Benefits and competitive implications

Strengthening platform integrity: Transparency obligations empower platforms to combat misinformation and harmful deepfakes. By integrating watermark detection and visible labels, platforms can reduce the spread of AI‑generated disinformation and rebuild user trust. This also helps regulators differentiate responsible platforms from those that allow unlabelled synthetic content.

Competition for trust and compliance capabilities: Platforms that invest in robust content authenticity infrastructure may attract advertisers seeking brand safety. Conversely, smaller or foreign platforms may face barriers due to compliance costs and potential fines, which could reduce competition. Partnerships with AI model providers could raise concerns about exclusivity and preferential access. The Commission’s competition policy brief warns that the concentration of key inputs—data, AI chips, cloud infrastructure and technical expertise—could lead to dependencies and foreclosurecompetition-policy.ec.europa.eu. Vertical partnerships could either provide access or entrench dominant players; regulators will monitor for self‑preferencing, bundling or exclusivitycompetition-policy.ec.europa.eu.

New opportunities for small platforms: Transparent labelling may allow niche platforms focused on authenticity to differentiate themselves. Start‑ups providing AI governance or moderation services could flourish as larger platforms outsource compliance functions.

5 Healthcare and medical devices

AI systems used for medical diagnosis, imaging and care recommendation are often high‑risk. The Johner Institute notes that AI systems intended for direct patient interaction—such as skin cancer diagnosis software or systems performing emotion recognition—require providers to inform users and document data use. Consequently, healthcare providers must disclose when patients interact with AI and label AI‑generated imagery or reports.

Benefits and competitive implications

Patient trust and informed consent: Transparent use of AI in diagnostic tools may alleviate patient concerns about automation and bias. High‑risk obligations (human oversight, risk management, high‑quality datasets) aim to ensure safety and reliabilitygoodwinlaw.com. Providers that comply may gain patient trust and reduce liability.

Innovation vs. compliance costs: Advanced AI medical devices require rigorous conformity assessments and documentation. Larger manufacturers may handle these requirements, whereas SMEs might struggle, risking reduced innovation. However, the Act mandates proportional fees and encourages Member States to reduce translation and assessment costs for SMEsartificialintelligenceact.eu. The limited number of “systemic‑risk” models (around 15 worldwide) means most AI tools used by smaller developers will face proportionate obligationsartificialintelligenceact.eu.

Competition among service providers: Healthcare institutions may prefer vendors with transparent AI to avoid reputational risk, thereby disadvantaging non‑compliant or opaque systems. On the other hand, early compliance may open access to EU markets for international AI healthtech firms.

6 Financial services

AI systems used for credit scoring, insurance underwriting, fraud detection or investment advice are considered high‑risk. Goodwin’s briefing notes that providers must implement robust risk and quality management, maintain technical documentation, ensure human oversight and pass conformity assessments before placing products on the marketgoodwinlaw.com. Deployers must monitor performance and inform individuals about decisions. Everest Group emphasises that limited‑risk systems like chatbots are subject to transparency obligations (disclosure and labelling), while high‑risk applications require human oversight and stringent controlseverestgrp.com.

Benefits and competitive implications

Building trust in financial decisions: Transparent communication about AI decision‑making helps customers understand why they were approved or declined for loans, addressing concerns about algorithmic bias and discrimination. Firms that provide clear explanations may attract socially conscious consumers and regulators.

Compliance as a competitive advantage: Global financial institutions that already follow rigorous governance frameworks (e.g., Basel III, GDPR) can leverage these systems to meet AI Act requirements. Compliance may become a prerequisite to access the EU market, potentially limiting non‑compliant foreign competitors. Service providers can offer AI governance, data management and large language model operations to financial institutionseverestgrp.com.

Barriers for smaller lenders: The need for robust documentation and human oversight could deter smaller fintechs, limiting innovation. However, by partnering with compliant technology providers or using regulatory sandboxes, SMEs can mitigate costs. The Commission encourages the development of standards and codes of practice to support industry adoptiondigital-strategy.ec.europa.eu.

7 Education and corporate training

Educational institutions increasingly use AI for personalised learning, assessments and administrative chatbots. These systems generally fall into the limited‑risk category unless they determine access to education or jobs (which would be high‑risk). Teachers must inform students when chatbots are used and label AI‑generated course materials or assessments. Deployers of emotion‑recognition tools (e.g., for monitoring attention) must inform students and cannot exploit vulnerabilities.

Benefits and competitive implications

Enhanced digital literacy: Transparency helps educators raise awareness about AI’s role and fosters digital literacy. A policy paper notes that building trust in content requires digital literacy, legal and regulatory frameworks, and international standardss41721.pcdn.co. Students who understand when AI is used can better evaluate content and develop critical thinking.

Trust and fairness: Transparent marking of AI‑generated educational content can reassure parents and regulators that assessments are fair and unbiased. Institutions that adopt open practices may attract students and funding. Conversely, high compliance costs could burden smaller schools; however, the voluntary nature of minimal‑risk obligations and the support for SMEs (proportional fees) mitigate this.

8 Gaming and virtual worlds

Games and virtual environments increasingly incorporate generative AI for character dialogues, world creation or NPC behaviours. These uses are generally limited‑risk but may venture into high‑risk if they involve gambling or financial transactions. RealityDefender’s article lists gaming and virtual worlds among sectors affected by deepfake labelling and distribution obligationsrealitydefender.com. Platforms must label synthetic content and ensure deepfakes are marked.

Benefits and competitive implications

Authenticity and moderation: Transparent labelling helps gamers distinguish between user‑generated content and AI‑generated scenarios, reducing confusion and preventing manipulation. It also assists in moderating harmful content (e.g., AI‑generated misinformation or harassment). Studios that incorporate authenticity standards may gain trust among players and regulators.

New creative tools and partnerships: Developers might partner with watermarking vendors or adopt open standards (e.g., C2PA) to ensure compliance. Early adopters can differentiate by offering players assurance about the authenticity of in‑game media. However, smaller studios may face higher compliance costs, although guidelines and a code of practice will provide claritydigital-strategy.ec.europa.eu.

9 Cross‑cutting issues and competition concerns

Concentration of key inputs and platform power

The European Commission’s competition policy brief warns that generative AI and virtual worlds are dominated by a few players controlling key inputs—large datasets, AI chips, computing infrastructure, cloud capacity and technical expertise—and that this concentration could lead to dependencies and foreclosurecompetition-policy.ec.europa.eu. Partnerships between large tech firms and AI developers provide distribution but may also raise abuse‑of‑dominance concerns (e.g., exclusivity, bundling, self‑preferencing)competition-policy.ec.europa.eu. For example, cloud providers could favour their own AI models or restrict competitors’ access to chips. Such dynamics might reduce competition in downstream verticals like marketing, media or healthcare.

The AI Act’s transparency requirements might intensify these concerns because compliance demands access to watermarking and detection technology. Large platforms may develop proprietary marking solutions and restrict access to others, further entrenching dominance. Regulators must monitor for anti‑competitive practices and ensure open standards.

Opportunities for new entrants and SMEs

Despite potential concentration, the Act creates new markets for compliance tools, auditing services and watermarking. Small companies offering detection, governance or AI explainability services can thrive. The Act also encourages open‑source models and smaller foundation models, which can reduce barriers and promote innovation【885248931030204†L424-L448】. SMEs can differentiate by emphasising ethical AI and transparency, and the Act’s proportional obligations and SME‑focused support measures (reduced fees, training) help mitigate burdensartificialintelligenceact.eu.

Conclusion and recommendations

The EU AI Act’s transparency obligations aim to make AI systems more trustworthy, mitigate misinformation and empower consumers. Generative‑AI applications in marketing, e‑commerce, media, entertainment, social media, education, gaming and other verticals are mainly limited‑risk. These sectors must disclose AI interactions and label AI‑generated or manipulated content from 2 August 2026artificialintelligenceact.eu. Deepfake labelling applies earlier, from 2 August 2025artificialintelligenceact.eu. High‑risk systems (financial services, healthcare, recruitment) face stricter requirements, including documentation, human oversight and conformity assessmentsgoodwinlaw.com.

Benefits: Transparent labelling can restore consumer trust, reduce misinformation, and create competitive differentiation. Customers are more likely to engage when brands are open about AI usage and maintain a human touchheygen.com. New markets for watermarking, detection and AI governance services will emerges41721.pcdn.co. The Act encourages ethical AI and invites SMEs to participate in standard settingartificialintelligenceact.eu.

Challenges: Compliance costs, especially for SMEs, risk stifling innovation and could entrench large incumbentsdigitalsme.eu. Unclear technical standards and missing regulatory sandboxes heighten legal uncertainty. The concentration of key AI inputs and vertical integration by major tech companies may raise competition concernscompetition-policy.ec.europa.eu. The Act’s extraterritorial scope means foreign platforms must comply if they reach EU usersrealitydefender.com.

Recommendations for verticals:

1. Develop clear disclosure protocols: Businesses should adopt standard phrases (e.g., “AI‑generated content”) and use internationally recognised metadata standards (e.g., C2PA) to tag synthetic media.
2. Invest in watermarking and detection technologies: Partner with providers such as Friend MTS or Google SynthID and implement machine‑readable markerss41721.pcdn.co.
3. Keep humans in the loop: Maintain human oversight for AI content creation and decisions; human review remains essential for compliance and qualityquibble.digital.
4. Train staff and communicate with users: Educate employees about AI obligations and inform users when AI is used; surveys show that consumer awareness improves engagementheygen.com.
5. Engage in standard‑setting and sandboxes: Participate in industry fora and regulatory sandboxes to influence practical standards and reduce uncertaintydigital-strategy.ec.europa.eu.
6. Monitor competition dynamics: Regulators and industry participants should watch for anti‑competitive behaviour such as exclusive partnerships or refusal to share watermarking technologies. Competition authorities may need to enforce pro‑competitive measurescompetition-policy.ec.europa.eu.

Overall, the AI Act’s transparency requirements will reshape vertical markets by making AI use more visible and accountable. Companies that embrace these obligations early, integrate authenticity tools and maintain human oversight can build trust and gain competitive advantage, while those that delay may face fines and reputational harm. The interplay between regulation and competition will determine whether the Act fosters a more trustworthy and innovative AI ecosystem or inadvertently concentrates power in the hands of a few dominant players.