Article 50 of the EU AI Act brings new transparency expectations to the customer-facing uses of AI.
As AI becomes more embedded across sales, marketing and service journeys, the next priority is ensuring customers are aware when it is used.
CX leaders should treat disclosure as part of experience design, ensuring it remains clear, timely and connected to a route to human support.
Colleen Jones, President at Content Science, explained to CX Today that transparency should be designed into the customer experience from the first interaction, rather than treated as a disclosure added after the fact.
“The test should be whether a reasonable customer can tell, without having to stop and investigate, that they’re interacting with AI," she said.
"Transparency works best when it is immediate, plain-language, and part of the experience, not buried in a policy.”
Article 50 Brings AI Disclosure Into Focus
Having taken effect on August 2nd, Article 50 was one of the first sections to become applicable, as this aims to solve the narrower and more immediate customer-facing issue: encountering AI.
Now, people must be informed when they are interacting with AI, requiring generated or manipulated content to be marked or labeled as these rules are intended to reduce deception and manipulation, and help people make informed choices.
Henna Virkkunen, Executive Vice-President for Tech Sovereignty, Security and Democracy, reinforced ethe importance of transparency as AI becomes more embedded in customer-facing experiences.
"With today's guidelines, the Commission supports the smooth and effective application of the AI Act to make AI systems interacting with people such as chatbots and AI agents and AI content more transparent and trustworthy."
"These guidelines support providers and deployers in meeting their obligations under the AI Act, while helping citizens know when they are interacting with AI."
As AI agents become more involved in customer service, sales, and marketing, this article becomes increasingly relevant to how brands build and maintain customer trust.
If customers feel they have been misled, brand confidence can suffer and potentially affect engagement, conversion, and retention, and clear disclosure can give customers the information they need to decide how much they trust an AI.
The New Disclosure Test for AI Sales Agents
Article 50 aims to bring transparency into the earliest stages of the buyer journey, requiring customer-facing AI system providers to inform them that they are interacting with AI, unless objectively obvious.
For sales and marketing teams, this article puts a spotlight on whether prospective customers can recognise its role before they rely on an interaction to decide.
Speaking with CX Today, Jessy Van Steenkiste, Senior Global Counsel for Regulatory Compliance Product, Privacy, and AI Governance at Parloa, argues that businesses should be cautious about relying on the exception for interactions where AI is considered obvious.
“Sales and marketing leaders should treat obvious as a low bar to lean on and a high bar to prove,” she stated.
For an AI sales agent handling the customer journey, assuming customers will recognize the technology themselves could create an unnecessary gap between a brand's expectations and its CX.
Furthermore, as AI agents become increasingly capable of replicating natural conversations, a customer may enter a 'human-appearing' conversation without realizing that an automated system is responsible for the responses.
“If your customer caller has to ask ‘wait, am I talking to a bot?’, you may have successfully rolled out a remarkably lifelike automated journey, but the moment for transparency and establishing trust already passed," she cautioned.
This requires disclosures being placed at the beginning of the interaction, whilst also providing a clear escalation route when a customer wants human assistance.
As Van Steenkiste warns:
“Name the AI, name the company, and offer a human within one line.”
The first disclosure can ultimately shape how customers interpret a conversation and how much confidence they place in the information received.
Deepfakes Raise the Stakes for Marketing Trust
Secondly, providers of generative AI systems must ensure synthetic audio, images, video and text outputs are marked in a machine-readable and detectable format, where technically feasible.
Furthermore, deployers now have obligations to disclose AI-generated or manipulated image, audio and video content that qualifies as a deepfake.
For sales and marketing teams, this creates an important distinction between using AI to assist content production and using it to create something that could make an audience believe a real person said, did or endorsed something that never happened.
In conversation with CX Today, Suvish Viswanathan, Head of Marketing at Zoho suggests businesses can establish practical internal boundaries around these uses.
“A practical and internal rule for sales and marketing is stock imagery and editing of a real photo or video shoot using AI steers clear of deepfake territory,” he said.
This allows teams to continue using AI for routine creative tasks while drawing greater scrutiny around fabricated content.
However, this risk becomes more significant when AI could infactually endorse a product, or a cloned executive voice makes a statement from a company leader.
Viswanathan emphasized:
“Where the line starts to be crossed is when AI starts to create voiceovers pretending to be people and fabricated customer testimonials.”
For sales and marketing leaders, this creates concern over authenticity and customer trust, as even where synthetic content may appear harmless, audiences may interpret it differently if they are unaware that AI was used to create it.
