This morning I read an interesting article by Rébecca Kleinberger, published on Beppe Grillo's blog.

In addition to being interesting in general, it addresses, even if only quickly, a point concerning ethical science that I did not know and that I would like to investigate further.

For many years, questions have been asked about the use that may be made by the producers of smart TVs of audio and / or video that these devices can and actually record and send to manufacturers.

A few years ago there was a flame on this subject, then it quickly passed into cavalry.

Today, however, more and more invasive devices are being added to these devices, the various Echo from Amazon, HomePod from Apple and Google Home from Google.

I, for my ignorance, I have always thought of the ethical implications linked to the fact that these devices, like the "old" smart TVs, listen to our words.

The article I read today, however, opened up a new world to me, a further level of problematic that, ignorance of mine, I did not know.

What Rébecca Kleinberger says on the subject:

For example: listening to very specific types of turbulence and the non-linearity of the voice can help to predict the early stages of Parkinson's, only through a phone call. Listening to the breath of the voice can help detect heart problems. And we also know that the change of rhythm in single words is an excellent indicator of depression. The voice is also very linked to hormone levels. Third parties who listen to female voices were able to accurately position the speaker in the menstrual cycle. Only with acoustic information. Now with technologies that are constantly listening to us, Alexa from Amazon Echo may be able to predict if you are pregnant even before you know it. Think about the ethical implications. The voice is also linked to how we create relationships. You have a different voice for every person you talk to. If I take a fragment of your voice and analyze it, I can know if you are talking to your mother, your brother, a friend or your boss. We can also use the voice posture to make predictions. That is, how you decide to position the voice when you talk to someone. The voice posture, when you talk to your partner, can also predict not only if, but also when you divorce. Rébecca Kleinberger

What the author of the article does, of course, are only examples and are only the current references of a new methodology of voice analysis, in the future things can only get worse / better according to the point of view.

Basically, therefore, not only these devices can perform actions or collect sensitive data that can be used for ethically questionable purposes, but they can do it even without necessarily understanding the words we speak, they can also analyze the tone, volume, pauses , etc. etc. of our voice and, through these analyzes, acquire even more sensitive data.

Let's think, for example, of a person who subscribes to a life insurance policy and the fact that the insurance, through one of the smart assistant manufacturers, can become aware of the fact that the subject is probably suffering from heart problems, is a smoker even if he declares otherwise, it could have Parkinson's disease, etc.

Let's think of a company that wants to hire a woman and might be informed that she is pregnant even BEFORE the woman herself is aware of it!

Or a manager who is in the throes of a depression, perhaps even unknowingly, and can be turned away by his company looking for information about him.

What about a law firm of marriage lawyers who call you proposing to protect you during a divorce case that you didn't even hypothesize?

Now, let's clarify, all of us, more or less, have the ability to understand many things about people only from the tone of the voice, from the pauses, etc. We understand when a man is attracted to a woman by the tone of his voice and vice versa; we can understand from the voice if a person has problems, etc.

Perhaps from the voice we can understand even if we face a smoker or not.

However, I don't think we would recognize a heart problem, a first stage of Parkinson's.

The ethical problem arises, as always, with technologies, not only for the fact that these capacities are wider than ours and in continuous rapid improvement, but the real vulnus is that they can acquire them quickly and pervasively, archive them indefinitely and spread them uncontrollably.

Contrary to what happens when we surf the Internet, when we write a text and when we pronounce particular words or formulate certain sentences, the central problem is that we do not have conscious control over the tone of the voice, the posture of the voice, etc.

For this reason the ethical implications are much greater, because we do not have control over how many and what information can be extrapolated from HOW we speak.

We can control what we say, which words we use, but not how we modulate the voice, a predominantly unconscious phenomenon.

Also in this case, as for many other ethical problems related to technology, it is really difficult to stem the problem and define the limits for data acquisition, for their diffusion and the controls that could eventually be done to verify that these poles are respected.

The first step, in any case, is the awareness, the acknowledgment that this problem exists and is extremely pressing and worrying.

How would you react if your request to order a pizza at home by Alexa answered you "Ok, the pizza will order it for you, but it wouldn't be better to save money since you're pregnant?"