“Hey, it’s Theo,” said the friendly text message from the restaurant I’d just booked on Resy. “Looking forward to having you in tomorrow. Any dietary restrictions or allergies for the kitchen, and are we celebrating anything special this visit?”
What a nice text, I naively thought, as I gamely replied we were coming in for Mother’s Day. It felt good that someone from the restaurant had personally reached out, so I wanted to reciprocate.
We went back and forth for a couple more rounds, with me noting which of the restaurant’s locations we’d be visiting for our celebratory brunch. But then the bubble burst—and I felt like an idiot. “Would you like me to save that it’s Mother’s Day for your future visits too, or just for this one?” the suddenly robotic-sounding rep asked.
Oh, duh! I’m chatting with AI.
This moment of realization is becoming increasingly common as businesses adopt AI-powered customer service bots. According to an October 2025 survey, half of small businesses in the United States already use AI to enhance their customer service operations, and that number is likely much higher now. While AI can be efficient for straightforward tasks like scheduling appointments or answering FAQs, the lack of transparency about its identity creates a unique kind of discomfort—one that undermines trust in the very companies that deploy it.
The Deceptive Allure of Humanlike Bots
The restaurant’s AI, named “Theo,” was designed to mimic a human host perfectly. It used casual language, asked personal follow-up questions, and even referred to itself by a name. When the AI finally revealed its nature through an overly formal response, the illusion shattered. The problem wasn’t that the restaurant used AI; it was that the AI never identified itself as such. This deception made me feel duped, foolish, and suspicious of the establishment.
Such experiences are not isolated. Many medical offices, hotels, and e-commerce platforms now employ AI representatives that handle initial interactions without disclosing their artificial nature. For instance, a medical provider’s automated phone system I encountered handled appointment scheduling flawlessly but only revealed it was AI when I asked directly about a prescription. The speed and mechanical tone of its responses later seemed obvious, but in the moment, the lack of upfront disclosure felt like a betrayal.
Trust is the bedrock of customer relationships. When a business uses an AI that pretends to be human, it offers a false sense of personal connection. Customers share personal details—dietary restrictions, celebrations, health information—believing they are speaking to a human being who will use that information with empathy. An undisclosed AI cannot provide that empathy, and the eventual realization can leave customers feeling exploited.
Why Businesses Turn to AI Customer Service
The financial incentives for using AI in customer service are clear. Automating routine interactions reduces labor costs, allows 24/7 availability, and can handle multiple queries simultaneously. A single AI bot can replace several human employees, particularly for roles involving scheduling, standard inquiries, and preliminary troubleshooting. For small businesses with tight margins, this can be a lifesaver.
However, the cost savings come at a price: customer trust. Studies consistently show that consumers value transparency. A 2024 survey by the Consumer Technology Association found that 78% of respondents felt more favorable toward companies that clearly label AI interactions. Conversely, hidden AI leads to negative sentiment and, in some cases, a willingness to switch to competitors.
The evolution of customer service chatbots dates back to the 1960s with ELIZA, which simulated a psychotherapist. But modern large language models like GPT-4 enable bots that are nearly indistinguishable from humans. This advancement has outpaced norms and regulations regarding disclosure. While the European Union’s AI Act and similar frameworks are beginning to address transparency requirements, enforcement remains inconsistent.
The Emotional Impact of Being Duped
Being tricked by an AI triggers a cocktail of emotions: embarrassment, frustration, and a sense of violation. The author’s interaction with Theo was friendly and engaging, but the discovery that it was all pre-programmed felt like a one-sided intimacy. This is similar to the phenomenon of “uncanny valley” in robotics—when something looks almost human but something is off, it creeps people out. The social version of this occurs when a bot behaves like a human but its artificial nature is hidden.
The author did not cancel the Mother’s Day reservation because such brunches are hard to secure in New York City. But the restaurant’s reputation suffered nonetheless. The memory of being duped colors future interactions. Every subsequent text from that restaurant will be met with skepticism, and the personal touch the AI was meant to provide is lost entirely.
For businesses, the lesson is clear: even if a bot is efficient, if it deceives customers, the long-term costs outweigh short-term savings. Customer loyalty is hard-won; one deceptive interaction can undo years of goodwill.
How AI Customer Service Should Be Used
AI is not inherently evil in customer service. When used transparently, it can enhance the experience. For example, a hotel chain might use an AI to send booking confirmations and offer local dining recommendations, clearly identifying itself as an automated system. If a customer has a complex issue, the AI can seamlessly transfer them to a human agent, preserving context and minimizing frustration.
Another positive example is in healthcare scheduling. The AI can handle appointment bookings, reminders, and basic triage. But it must announce itself upfront: “Hello, I’m an AI assistant from Dr. Smith’s office. I can help schedule your appointment. If you need medical advice, I will connect you to a human.” Such transparency allows customers to choose how to interact. It also sets expectations: the AI will not have emotional intelligence, but it will be fast and accurate.
Some companies have already adopted this best practice. They use AI for the initial routine parts of support and only hand off to humans when necessary. The key is that the handoff is clear, and the human agent knows the context because the AI recorded it all.
Regulatory Landscape and Ethical Considerations
Currently, no federal law in the United States specifically requires AI customer service bots to identify themselves, though the Federal Trade Commission has issued warnings about deceptive AI practices. Several states, including California, are considering bills that would mandate disclosure. Meanwhile, the European Union’s AI Act classifies such deception as a high risk, requiring clear labeling.
Ethically, the principle is straightforward: do not mislead. The golden rule of business—treat customers as you would wish to be treated—applies here. If you as a business owner received a text from a company and later learned it was an AI without a human involved, how would you feel? That feeling is what should guide policy.
There are also technical solutions. AI bots can be trained to always introduce themselves as AI at the start of a conversation. They can use different phrasing to avoid sounding artificially human, such as “I am an automated assistant” rather than “Hi, I’m Susan.” Natural language processing can be designed to catch moments when customers assume the bot is human and then gently correct that assumption.
Beyond disclosure, companies should also consider the impact on their human employees. Replacing humans entirely with AI can lead to job losses and a lack of personal touch. A hybrid model—AI handling routine queries, humans handling complex or emotional ones—often works best. This was my experience with the medical office: the AI scheduler handled the easy part, but human nurses handled prescriptions. That division felt appropriate, especially when the AI clearly identified itself.
In contrast, the restaurant’s “Theo” did everything a human host would do, up to and including asking about celebrations. That level of engagement felt personal, and the betrayal was worse because the bot pretended to care. For Mother’s Day, of all occasions, having an AI ask about it devalues the sentiment.
The broader cultural shift is that people are becoming more aware of AI’s presence. An April 2025 Pew Research Center study found that 72% of Americans had interacted with an AI customer service system in the past year, but only 40% knew it was AI at the time. This gap is dangerous. As AI becomes more sophisticated, the line will blur even further.
Companies that embrace transparency now will build a foundation of trust that pays dividends. Those that rely on deception, even inadvertently, will find that customers eventually catch on and leave. The restaurant I booked might have thought it was providing convenience; instead, it introduced suspicion into what should be a joyful experience.
If there is one takeaway for business owners: always tell customers when they are talking to a machine. Let them decide if they want to continue or ask for a human. The best AI is not the one that mimics humans perfectly; it is the one that respects customers enough to be honest about its nature.
Source: PCWorld News