“AI should not pretend to be human”

Written by Andrea Schmits
Photos by Stefan Jermann

March 05, 2020

Speak instead of typing – nowadays, communicating with computers often involves language bots. In his book “Dialogue Robots”, Armin Sieber analyzes the widespread dissemination of these bots. In this interview he describes the advantages – but also urges policymakers and the public to be aware of the dangers.

“AI should not pretend to be human”

Mr. Sieber, are we already engaging in dialogue with robots without realizing it?

This is not very likely today. One reads again and again about machines that order pizzas or make appointments over the phone. Such systems are generally not suitable to serve the masses. Certainly, dialogue robots are already being used extensively, for example as language assistants, such as Alexa and Siri, or as e-commerce bots by banks or travel agencies. But rather than for conventional dialogue, we use these bots to mostly give commands. After all, it is very difficult to program a machine so that it is able to conduct a credible conversation in terms of emotionality and voice response, as humans do. However: I also do not believe that this is the right way to go.

What exactly do you mean by that?

The requirement for artificial intelligence to appear as human as possible is fundamentally flawed. A human being has a physical body, a history, and has experienced success or failure. A machine will never be human. At best it can simulate a human being. But why should we want it to do so? Artificial intelligence should remain just that – and make it known what it is – and it should not pretend to be a person. The more this were to happen, the more we would run into problems of authenticity in the public discourse. However, dialogue robots offer society other benefits.

For example?

Dialogue robots are an interface that makes it easier for us to access the digital data space with all its knowledge resources. They scour the Internet for information for us, present the weather forecast, or remind us of upcoming meetings. They also enable us to conveniently and efficiently remotely control devices such as the TV, the stove, or the heating using voice commands. Dialogue robots are also of great interest to companies that use them in marketing or customer service situations, for example to greet customers, answer questions, or provide help in online shops.

Will such dialogue robots soon be a must-have for companies?

There is hardly a big-brand company that is not already looking into making its products dialogue-enabled. “Voice-first” is the keyword here. In this regard, Europe is not yet as advanced as the US, China, or Southeast Asian countries, for example. In the meantime, there is an ever-increasing number of everyday products that are networked or that can be controlled using a natural language interface. This trend will remain, even if the media hype about bots has just subsided a bit.

In your book you express the concern that “voice first” could mean the end of written text.

Yes, I view this as a serious threat. We are currently experiencing a dialogue revolution: An increasing proportion of mass media communications, which have until now been based on written text, may become dialogue-based in the future. However, even replacing just one quarter of today’s text-based communications with voice-based applications could result in significant socio-cultural shifts. Writing not only conveys information, its function is to structure it - and thus becomes a tool of insight. Therefore, in my opinion, an efficient culture of knowledge is barely conceivable without writing. However, it must also be emphasized that the automation of dialogue does not necessarily have to result in the displacement of writing. Chats are still text-based. And voice-based interfaces, such as those used on PCs, can also be used as an interesting addition to text processing.

Are we underestimating the risks?

Maybe. The topic of dialogue robots is like an iceberg: Many people only see the tip consisting of cool machines. But below the surface, there is a whole mountain of hidden problems. This is a topic that policymakers and the public need to think about. The developments are more profound than we may realize.

The book: Dialogue robots

The book "Dialog Robots" by Armin Sieber: How bots and artificial intelligence change media and mass communication

How bots and artificial intelligence are transforming the media landscape and mass communications.
(Dialogroboter: Wie Bots und künstliche Intelligenz Medien und Massenkommunikation verändern.)

Armin Sieber, Springer-Verlag

* This book is only available in German for the time being.

The author

Portrait of Armin Sieber

Armin Sieber is a research assistant at the University of Regensburg’s Department of Media Studies. Previously, he worked as a communications manager for a number of companies and has been running his own corporate and crisis communications consulting firm since 2015.


Artificial intelligence, autonomous driving, smart homes: The future is closer than we think. As a journalist, I have a passion for many topics – and I enjoy telling readers the stories behind them. I have been doing this for print and online magazines, newspapers, and blogs for more than a decade. Before that, I studied journalism and sociology at the University of Zurich.
Taught to drive by my grandfather at the age of eight. Hurtling down every set of steps on my BMX by the age of ten; racing down every black diamond slope on my snowboard when I was fourteen. I am fascinated by people who have a story, and I want to tell it. My prime focus is on the relationship between human and machine, artificial intelligence, and finding inner peace in a world full of noise.

Read more