Here is an illustrative scenario. It is not a real customer story.
A customer notices a wrong charge on her bill. She opens the company’s chat window. A friendly assistant greets her and asks how it can help.
She explains the problem. The assistant offers three help articles. She asks for a person. It offers the same three articles again.
After twenty minutes she is still stuck. She starts to wonder about something bigger than the bill. Is the company even allowed to do this? Do I have a right to talk to a human?
The honest answer is more complicated than yes or no. This article explains what EU rules say, what they do not say, and what you can do today.
It is general information, not legal advice. The rules are changing, so check the official sources linked at the end.
Some context first. The EU AI Act’s chatbot disclosure rule has applied since 2 August 2026, according to Aptus Legal. That rule is about telling you that you are talking to AI. It is not about giving you a person.
That gap is why companies are rethinking how chatbots and people share the work. Providers such as FlairsTech describe a hybrid approach on their AI customer experience services page. AI handles repetitive work, and human specialists handle empathy and problem-solving.
The rest of this article looks at what the rules require, what they do not, and what good service looks like in practice.
Key takeaways
- There is no single EU right to a human in every chat. No general rule covers every chatbot in every sector.
- Some rights do exist in specific cases. They cover automated decisions with serious effects and some financial services.
- Companies must keep other ways to reach them. A chatbot can be an extra channel. It cannot be the only one.
- The EU AI Act gives you a right to know. Many chatbots must tell you that you are talking to AI.
- More rules may be coming. A Digital Fairness Act is expected, but its content is not yet known.
Is there a general right to talk to a human?
Short answer: not in general. But several rules help in specific situations.
Many people assume the EU guarantees a human on the line. It does not, at least not across the board. Legal commentators say there is no general cross-sector right to human contact with a chatbot.
That does not mean you are powerless. Three kinds of rules matter.
1. Decisions made only by machines
The General Data Protection Regulation, or GDPR, protects people from some decisions made solely by automated processing. The protection applies when a decision has legal effects or similarly significant effects on you.
Where such decisions are allowed, safeguards apply. They include a way to ask for human review. Think of a refused application or an account closed by an automated system.
This is about decisions, not conversations. A chatbot that answers a question about opening hours does not trigger it. A system that decides whether you get a loan might.
2. Financial services
Some sector rules go further. According to a law-firm summary, financial services rules such as the Consumer Credit Directive give consumers a right to human intervention. This applies when they deal with automated systems like chatbots.
If your problem involves credit or similar financial products, ask for human intervention and mention your rights. The exact wording of those rights is in the legal texts, so read the official source if you need to rely on it.
3. Other ways to contact the company
EU consumer rules also expect businesses to offer ordinary contact channels, such as a phone number, email address or postal address. The same summary says a chatbot cannot replace them. It cites a European Court of Justice case on the principle that consumers must be able to contact a business directly, quickly and effectively.
So a company may add a chatbot. It should not make the chatbot the only way in.
Put together, these rules give you real tools. They do not give you a guaranteed human in every chat window.
What does the EU AI Act say about chatbots?
Short answer: it says you have a right to know that you are talking to a machine.
The EU AI Act is the bloc’s main law on artificial intelligence. Article 50 sets transparency rules. For chatbots, the idea is simple. People should be told when they are dealing with AI.
According to Aptus Legal, the chatbot disclosure duty has applied since 2 August 2026. The Digital Omnibus did not change that date. The same source says the notice should come at the start of the conversation, in a form users will notice.
There is an exception where it is obvious you are dealing with AI. Regulators expect providers to use that exception sparingly.
Two points are easy to miss.
- It is about disclosure, not access. Article 50 does not give you a right to a human. It gives you a right to know.
- Human oversight rules are narrower. The Act’s human oversight duties target certain high-risk systems, not every customer chatbot.
A company that brands a chatbot built on someone else’s model may count as its provider. That means the duties could fall on the company, not only on the model’s developer.
Five common myths about chatbots and your rights
Confusion is common. Here are five myths worth clearing up.
Myth 1: “The EU guarantees a human in every chat.” It does not. Legal commentators say there is no general cross-sector right to human contact with a chatbot. Rights exist in specific cases.
Myth 2: “A chatbot is illegal if it cannot transfer me.” Not necessarily. A chatbot may be one channel among several. Problems arise when it is the only way to reach the company, or when a rule gives you a right to human intervention.
Myth 3: “The AI Act gives me a right to speak to a person.” Article 50 is about disclosure. It aims to make sure you know you are dealing with AI.
Myth 4: “If I am talking to a bot, nothing I say counts.” What you write in a chat can be useful evidence. Keep a record of it.
Myth 5: “New EU rules will fix this soon.” A Digital Fairness Act is expected, but nobody knows yet whether it will cover chatbots. Do not count on it.
The common thread is simple. The law offers some protection, but it is patchy. Good service fills the gaps.
What does a good route to a human look like?
Short answer: you reach a person quickly, and that person already knows what happened.
The law sets a floor. Good service goes beyond it. Whether or not a rule applies, good customer service shares the same features.
- A visible way out. A button or a simple word, such as “agent”, reaches a person.
- Context that travels. The person who takes over can see what you already said.
- No repeating yourself. You do not retell the story or verify everything again.
- Your own language, kept throughout. This matters in a region with many languages.
- Honesty. When the AI does not understand, it says so and hands over.
The practical model is hybrid. AI handles repetitive work, such as lookups and routine questions. Trained people handle empathy, judgement and complex cases.
A hybrid model is only as good as its handover rules. Whoever runs the service should decide, in advance, when the AI must stop and a person must step in.
What may change next?
The European Commission is preparing a Digital Fairness Act. It aims to strengthen consumer protection online.
Many observers have asked whether it might deal with chatbots and human contact. Legal commentators have said it could address the gap. Nobody knows yet.
According to a European Parliament research briefing dated October 2026, the Act has not yet been proposed. It is expected in the fourth quarter of 2026. The briefing does not say the Act will cover chatbots or a right to human contact.
Its closest topic is automated contracting, where the Commission’s review found it needed more evidence before drawing firm conclusions.
So the right way to read this is simple. A change is possible. It is not promised.
How do these rules apply in everyday situations?
Rules are easier to understand with examples. The cases below are illustrative. They show how to think about each situation. They are not legal conclusions.
A wrong charge on a phone or utility bill
No special human-intervention rule is obvious here. But you can still insist on ordinary contact channels. Ask for a phone number or email address. Keep records of the dispute.
A loan or credit application refused by an automated system
This is closer to the protected cases. A decision with significant effects on you, made solely by automated means, can trigger GDPR safeguards. Financial services rules may add a right to human intervention. Ask for a person to review the decision and say why.
An account closed or blocked by an automated check
If the closure has serious effects on you, ask for human review and an explanation. Describe the impact clearly. Mention the automated nature of the decision.
A delivery that never arrived
Use the chatbot first if it is quick. If it stalls, move to another channel. Ask for a reference number. Escalate in writing.
A chat that never says it is AI
Under the AI Act’s disclosure rule, many chatbots should tell you at the start. If a service pretends to be a person, you can report it to a consumer authority. Save screenshots first.
In all five cases, the same habits help. Stay calm. Keep records. Ask for a person. Use more than one channel.
What can you do today?
You cannot rewrite the rules. You can make your case easier to solve.
- Ask clearly. Type “agent” or “human”. If that fails, say you want to speak to a person.
- Keep a record. Note the date, the channel and any reference number. Save screenshots.
- Use another channel. Try the phone or email if chat stalls. Companies should offer them.
- Ask for human review of automated decisions. If a machine made a decision with serious effects, say so and ask for a person to review it.
- Mention financial services rules if they apply. If the issue involves credit, ask for human intervention.
- Escalate to a consumer body. If the company does not respond, contact a consumer authority. In Finland, the Finnish Competition and Consumer Authority (KKV) provides consumer advice. Check its current contact routes first.
None of this is a guarantee. It does give the next person something to work with.
What should companies do?
Companies that use chatbots can build trust with a few simple steps.
- Say it is AI. Tell people at the start. Make the notice easy to see.
- Offer a clear route to a person. Do not hide it behind loops of repeated answers.
- Keep traditional channels. Phone, email and post still matter.
- Write down handover rules. Define when the AI must stop.
- Measure resolution. A closed chat is not a solved problem.
- Check the legal texts. Rules differ by sector and are still changing. Ask a qualified lawyer where it matters.
Each step is simple. Together they decide whether customers feel served or stuck.
Frequently asked questions
Can a company force me to use only a chatbot?
Per legal commentary, EU consumer rules expect traditional contact channels such as phone or email. A chatbot may be added. It should not replace them.
Does the company have to tell me I am talking to AI?
Under the AI Act’s Article 50, many chatbots must. There is an exception where it is obvious. Regulators expect that exception to be used sparingly.
Can I demand that a person review an automated decision?
Where a decision is made solely by automated means and has legal or similarly significant effects, GDPR safeguards apply. They include a way to ask for human review. Details depend on the case.
Does this apply outside the EU?
The EU rules apply to services offered in the EU, whatever the company’s location. Check the official texts for the exact scope.
What if the company ignores my request?
Keep a record, escalate in writing and contact a consumer body. In Finland, the Finnish Competition and Consumer Authority (KKV) provides consumer advice. Check its current contact routes.
Will the Digital Fairness Act give me a right to a human?
Nobody can say yet. The Act has not been proposed. Even once it is, its content will take time to settle.
A short glossary
- AI Act. The EU’s main law on artificial intelligence.
- Article 50. The part of the AI Act that sets transparency rules, including for chatbots.
- GDPR. The General Data Protection Regulation. It covers personal data and some automated decisions.
- Automated decision. A decision made without human involvement.
- Human intervention. A person reviewing or changing a decision.
- Hybrid support. A service where AI and people share the work.
- Digital Fairness Act. An expected EU proposal on consumer protection online.
The bottom line
You may not always have a legal right to a human in a chatbot conversation. You do have a growing right to know when you are talking to a machine.
In some cases, such as serious automated decisions and certain financial services, you can ask for a person to step in.
Companies that make it easy to reach a person will keep more customers. They will also be better prepared for the rules that may come next.
