A recent Swiss study secretly introduced AI-based chatbots into a public online debate forum, without the participants’ knowledge.
The goal was simple: to see whether artificial intelligence could successfully persuade humans without them realizing they weren’t talking to another person.
The result? Yes, it often could.
Even the forum moderators were unaware that AI agents were participating. The experiment has since sparked not only scientific interest, but legal and ethical debates as well.
This was the topic of discussion between Kovács-Nagy Áron and Levente Szabados, associate professor at the Frankfurt School of Finance & Management and lead advisor at Neuron Solutions.
Can AI Really Influence Us?
Absolutely.
According to Levente Szabados, AI doesn’t just convey information, it can shape opinions and emotions, which carries serious societal implications.
To understand this, we need to acknowledge that human beings influence each other through language.
We are a cooperative, social species. We talk, argue, share ideas, and change each other’s minds through words.
Modern large language models (LLMs) are now capable of generating complex, nuanced reasoning, tailored to the listener. The result can feel remarkably human, and highly persuasive.
This influence can stir emotions and shift perspectives, which is why major AI developers are investing heavily in defining the boundaries of such systems and ensuring their use remains within ethical limits.
Can the Impact of AI persuasion Be Measured?
That’s one of the biggest challenges.
Even in traditional human conversations, measuring persuasion and opinion change is notoriously difficult. If I think differently today than I did yesterday, was it because someone persuaded me, or did I just reflect on something?
Levente Szabados explains that researchers are currently working to develop more scalable and accurate methods to track how opinions shift, particularly under AI influence.
This kind of research is essential if we want to use AI responsibly in communication, but we’re not there yet in terms of precision or safeguards.
Can Chatbots Be Ethical in Social Media?
Yes, but only if we’re clear on what their purpose is, and how they impact us.
Every form of communication has an intent, whether it comes from a person or a machine.
If someone says, “Nice weather today,” that’s already a subjective statement. An AI can generate similar messages, and even build full arguments.
So the real question isn’t whether AI should be allowed to speak in public forums, but rather:
How do we respond to what it says?
Who has the right to represent public opinion?
Where’s the line between communication and propaganda?
These questions have always existed, but AI is now forcing us to answer them with greater urgency.
🎧 Listen to the full conversation with Levente Szabados by clicking here.
