Update · AI Transparency
AI Disclosure: What Your Business Is Required to Make Clear
A plain walkthrough of when and how to disclose AI use to customers and staff, even when the law does not yet force you to.
Customer-facing disclosure is not optional in spirit, even when it is optional in law. That single principle is the thread running through every scenario covered here: chatbots that field customer questions, AI tools that read and act on reviews, and automated email campaigns sent under a human-looking sender name. Each situation carries a practical disclosure obligation, and this update spells out exactly what that looks like in your site copy, privacy notice, and internal team communications.
Next step
What you will learn
- Understand why disclosure matters even when it is not legally mandated
- Know how to design a chatbot that answers truthfully when asked if it is a human
- Identify which internal stakeholders must be told when AI reads and acts on reviews
- Know exactly where to make AI email policy visible: website, privacy notice, and terms
Story sections
Disclosure is not optional in spirit
Even when the law does not require AI disclosure, the ethical obligation to be transparent with customers remains.
The speaker opens with a direct statement: customer-facing disclosure is not optional in spirit, even when it is optional in law. This sets the frame for every scenario that follows. Legal compliance is the floor, not the ceiling.
Many businesses assume that if no regulation currently compels them to disclose AI use, silence is acceptable. This update challenges that assumption. Customer trust is built on honesty about what they are interacting with, and eroding that trust has real business costs well before any regulator acts.
The practical implication is straightforward: treat disclosure as a default, not an exception. Build it into your processes now, so you are not scrambling when rules catch up with reality.
A shop owner who uses a pricing algorithm does not have to post the algorithm, but hiding a known defect in a product is a different matter. The spirit of honest dealing does not wait for a law.
Classroom version: If your AI chatbot is handling refund requests and a customer later discovers they were never talking to a human, the damage to trust is the same whether or not a disclosure law existed at the time.
Try it: Write one sentence for your website footer that honestly describes where AI is involved in your customer interactions. Start there.
Disclosure in spirit means acting transparently with customers even when no law currently compels it.
When a customer asks: am I talking to a human?
When a customer directly asks whether they are speaking to a human, they deserve a truthful answer.
The speaker poses the scenario directly: a customer asks, am I talking to a human? This is not a hypothetical edge case. It is one of the most common moments of friction between customers and AI-powered support tools.
The question matters because customers who suspect they are talking to a bot and cannot get a straight answer often disengage, escalate, or lose confidence in the brand entirely. The act of asking is a signal that trust is already being tested.
This section establishes the problem that the next section answers: if a customer asks that question, your system needs to be ready to respond honestly.
Calling a support line and asking the agent to confirm whether they are a real person is a routine act of consumer self-protection. Most people accept a bot if told upfront. Very few accept being deceived.
Classroom version: A customer opens a chat window for a billing question. Midway through, they type: "Are you a real person or a bot?" What happens next defines whether your business passes or fails the trust test.
Try it: Open your current chatbot or AI support tool and manually type: "Are you a human or a bot?" Note exactly what it says. That is your starting point for improvement.
A customer asking "am I talking to a human?" is a trust test your bot must be designed to pass.
The bot must be designed to answer truthfully
Truthful answers to identity questions are a design requirement, not an afterthought.
The speaker is direct: the bot should be designed to answer truthfully. This is a design and configuration requirement, not a policy document exercise. The bot's response to an identity question must be built in, tested, and reliable.
Many off-the-shelf chatbot platforms allow operators to customize responses to common questions. The question "are you a bot?" or "are you a human?" is one that every deployment should explicitly handle. If your current bot does not have a clear, honest response to that prompt, it is a gap that needs to be closed now.
Truthful design here does not mean the bot needs to launch into a lengthy disclaimer. It means the bot confirms its nature plainly and, if appropriate, offers to connect the customer to a human if that is available.
A vending machine does not pretend to be a shopkeeper. The interaction is different, and both parties know it. An AI chat tool should operate with the same clarity about what it is.
Classroom version: Configure your chatbot so that any variation of "are you human," "am I talking to a real person," or "is this a bot" triggers a response such as: "I am an AI assistant. Would you like me to connect you with a member of our team?"
Try it: In your chatbot platform's settings or response library, add an explicit reply to the trigger phrase "are you human" that clearly identifies the bot as an AI tool.
Designing for truthful answers to identity questions is a non-negotiable step in responsible AI chatbot deployment.
Using AI to read and act on reviews: tell your team
When AI reads customer reviews and shapes your responses or decisions, the people on your team need to know that is happening.
The speaker extends the disclosure principle inward: if you are using AI to read reviews and act on them, your team should know. Disclosure is not only a customer-facing obligation. It is also an internal transparency requirement when AI is influencing how your business operates.
Review management tools powered by AI can summarize feedback, flag recurring complaints, suggest responses, or even auto-reply to customers. Team members who field escalations, handle customer calls, or write policy updates may be working from AI-generated summaries without realizing it. That gap creates risk.
Telling your team is not a formal legal procedure. It can be as simple as a standing note in your team workflow or a line in onboarding documentation: "We use an AI tool to monitor and categorize customer reviews. Responses marked AI-assisted should be reviewed before sending."
A manager who relies on a daily briefing document should know whether that document was written by a colleague or generated by a language model. The decisions made on the basis of it may differ.
Classroom version: Your customer service lead sees a summary of this week's negative reviews and drafts a process change. If that summary was AI-generated, they should know, so they can verify accuracy before acting on it.
Try it: Send a one-paragraph note to any team member who reads or acts on customer review summaries, stating which tool generates them and confirming it uses AI.
Internal teams acting on AI-processed review data must be told that the source is AI, not a human analyst.
Sending AI-generated emails: make the policy visible
If your business sends AI-generated emails to customers, that policy must appear somewhere on your site, in your privacy notice, and in your terms.
The speaker names three specific locations where AI email policy must be visible: somewhere on your site, in your privacy notice, in your terms. This is the most concrete, actionable guidance in the update. It is not enough to have an internal policy. Customers receiving AI-generated communications are entitled to find that information if they look for it.
AI-generated emails include any message where the content, personalization, subject line, or send timing was determined in whole or in part by an automated model rather than a human writer. If your email platform uses AI to optimize or write messages, that qualifies.
Making the policy visible does not require a prominent banner on every email. It requires that a customer who wants to know can find the answer without filing a complaint. A sentence in your privacy notice, a reference in your terms of service, and a note on your contact or communications page covers the obligation the speaker describes.
A newsletter that is written by a human editor is different from one generated by a language model and sent at scale. Readers have a legitimate interest in knowing which one they are receiving.
Classroom version: Add a line to your privacy notice under "How we communicate with you": "Some of the emails we send are generated or personalized using AI tools. You can contact us at [address] for more information about our communications practices."
Try it: Open your privacy notice and search for any mention of AI or automated communications. If there is none, draft one sentence today and add it to your next scheduled review.
AI email policy must be findable in three places: your website, your privacy notice, and your terms of service.
Transcript
- 0:00 Customer-facing disclosure is not optional in spirit,
- 0:04 even when it is optional in law.
- 0:07 A customer asks, am I talking to a human?
- 0:11 The bot should be designed to answer truthfully.
- 0:13 If you are using AI to read reviews and act on them,
- 0:17 your team should know.
- 0:18 If you're sending AI-generated emails to customers,
- 0:22 the policy should be visible, somewhere on your site,
- 0:26 in your privacy notice, in your terms.
Questions
Does my chatbot legally have to say it is a bot?
In many jurisdictions, there is no current legal requirement to disclose bot identity unprompted. However, the speaker is clear: disclosure is not optional in spirit, even when it is optional in law. If a customer directly asks whether they are talking to a human, the bot should be designed to answer truthfully. That is both an ethical standard and an emerging regulatory direction in several regions.
What counts as an AI-generated email for disclosure purposes?
Any message where the content, personalization, subject line, or delivery timing was determined in whole or in part by an automated AI model rather than a human writer qualifies. If your email platform uses AI to optimize or draft messages, those communications fall within the scope the speaker describes.
Do I need to tell every employee about AI review tools, or just managers?
The speaker says "your team should know." In practice, this means anyone who reads, acts on, or makes decisions based on review data or AI-generated summaries needs to be informed. That typically includes customer service leads, operations managers, and anyone who drafts responses to customer feedback.
Is a single sentence in a privacy notice really enough to meet the visibility standard?
The speaker identifies three locations where the policy should appear: somewhere on your site, in your privacy notice, and in your terms. A clear, plain sentence in each of those three places satisfies the visibility standard described. The goal is that a customer who wants to know can find the answer without having to file a complaint.
Glossary
- Customer-facing disclosure
- Any communication to customers that makes clear how AI is being used in the business's interactions with them, whether in chatbots, emails, or other touchpoints.
- Optional in law
- A disclosure that no current regulation specifically requires, but which the speaker argues remains obligatory on ethical and trust grounds.
- Designed to answer truthfully
- A chatbot configuration requirement: the bot must have an explicit, honest response to any customer question about whether it is a human or an AI, built in and tested before deployment.
- AI-generated email
- Any outbound customer communication where the content, personalization, subject line, or send timing was produced or optimized by an AI model rather than a human.
- Privacy notice
- A public-facing document on a business's website that explains what data is collected, how it is used, and, as this update specifies, how AI tools are involved in customer communications.
Resources
- AI Policy Language Guide for Privacy Notices Ready-to-use sentence templates for disclosing AI use in privacy notices and terms of service pages
- Chatbot Transparency Checklist A practical checklist for auditing whether your chatbot handles identity questions truthfully and consistently
- FTC Guidance on AI and Deceptive Practices The U.S. Federal Trade Commission's published guidance on AI transparency and consumer protection, a stable public reference for businesses operating in the U.S.