display
Kenza Ait Si Abbou, born in Fès in north-eastern Morocco, studied electrical engineering in Valencia and industrial engineering in Germany.
Today, as a manager for robotics and artificial intelligence at Telekom, she is responsible for process automation, among other things.
In addition to her mother tongues Arabic and French, the 40-year-old speaks fluent Spanish and Catalan, Chinese and Japanese and of course German.
Ait Si Abbou, who lives in Berlin, has received multiple awards, including the “Digital Female Leader Award 2018”, as a book author: In “Don't panic, it's just technology”, she passionately advocates more diversity in the IT industry and one more relaxed handling of artificial intelligence.
WORLD:
Ms. Ait Si Abbou, the corona crisis has lifted the veil over a country that is in need of redevelopment: over-bureaucratic, disorganized, poorly managed and in offices, authorities and schools largely at the technical level of the past century.
What do you say about this debacle?
display
Kenza Ait Si Abbou:
It is actually the case that the Corona crisis shows the
undesirable
developments.
It is a shame when a country like Germany first has to get into such a crisis in order to tackle digitization in education and health care, in fact to even notice its shortcomings in these areas.
For me personally, relatively little has changed in the Corona crisis: I've been working mostly in the home office for a long time.
I don't need more than a laptop and a cell phone.
My children still go to daycare and when it closed it was of course a huge challenge.
But at least I didn't have to take on home schooling like many parents of school children.
WORLD:
To contain the pandemic, the state uses the measures and methods of the Middle Ages: the isolation and isolation of the infected.
Technical aids are only reluctantly used, sometimes not even considered.
In this scenario, does fear mix with complacency and risk aversion to create a dangerous backwardness?
Ait Si Abbou:
I would put it a little more positively: You want to plan everything down to the last detail.
In Germany there is a will for perfectionism, at the same time there is a lack of flexibility to be able to react to new situations.
You want to know in advance what is coming and plan a reaction.
But in crisis situations like this you just have to have courage and act.
But yes, there is already a great need for security and planning in Germany.
display
WORLD:
Your book “Don't panic - it's just technology” also revolves around the suspicion with which many people view artificial intelligence in particular.
What made you want to write this book?
Ait Si Abbou: In
recent years I have noticed that the topic of artificial intelligence is very fearful.
Many people associate AI with the terminator or with a technology or something magical that cannot be controlled.
And that of course fuels fears.
However, I have repeatedly made the experience that people lose their distrust of AI if you explain the nature and methods of AI to them in simple words.
And this understanding of AI is very important because technology is changing the lives of all of us.
That is why it is so important to understand how it works.
And because I especially want to reach those people who are still analog, so to speak, I decided to use the traditional medium of the book.
It is intended to help demystify AI.
display
WELT:
In your book you also plead for more diversity in the digital economy.
For example, there is a chronic shortage of women in the IT industry.
Why is that, a lack of interest, a lack of talent?
Ait Si Abbou:
From my point of view, it's a question of socialization.
There are no studies to prove that it has anything to do with biology or genes.
It's definitely socialization.
You can see this, for example, in the Eastern Bloc, where about as many men as women work in technical professions.
In Morocco, where I grew up, girls and boys also show the same talents when it comes to math skills, for example.
WORLD:
Why are comparatively few women drawn into the tech milieu?
Ait Si Abbou:
This is also due to the image of these Techi programmers, for example: They are all supposedly totally closed and unsympathetic, wear hoodies, have acne and just sit in front of the computer all day.
This scares off many girls of puberty.
And so they choose another profession.
WORLD:
According to the general opinion, there is an uncomfortable macho culture in Silicon Valley.
Ait Si Abbou:
The phenomenon seems to exist.
I didn't experience it personally, but it was taken from books like “Brotopia” by Emily Chang, or media coverage.
Even if many women are fighting against it, the IT industry remains in the perception of zero attractive for girls.
WORLD:
The scene is almost 90 percent dominated by “young white men”, as you say.
This also has consequences in that these young white men also program the prejudices and errors of young white men into their AI systems.
display
Ait Si Abbou:
Right.
Computers themselves have no consciousness, they can only make their decisions based on the logic that we humans have given them.
Let me give you an example: There are large companies that leave the pre-selection of the applications received for advertised positions to AI programs.
You do this on the basis of objective criteria.
And yet there are prejudices or so-called distortions that humans bring to machines: if in the past, for example, we only hired white male graduates from the Technical University of Munich for a management position, the machine will learn that people with this qualification are particularly suitable.
WORLD:
With the result that the hiring manager is hiring more and more people who are similar to himself and all the others he has already hired.
Does this result in simplicity instead of diversity?
Ait Si Abbou:
These subconscious self-reinforcing prejudices are indeed a problem.
It is not as if a software engineer sat down and programmed a “graduate of the Technical University of Munich” as a recruitment criterion.
But the decisions we made in the past are passed on to the machine.
And that is the reason why I advocate diversity in the development teams.
We have to train people and educate them about their subconscious prejudices, but at the same time we have to work to ensure that all teams are mixed.
WORLD:
AI systems are prejudicial and also tend to discriminate.
Almost every third dark-skinned woman is not recognized by facial recognition programs.
Is AI Racist?
Ait Si Abbou:
That is structural racism.
The images with which AI systems are fed or trained are primarily drawn from old databases that are still mostly filled with photos of white men.
The ideals of beauty in the world also have a reinforcing effect.
This means that even if we diversify the historical databases - and many initiatives, associations and universities are working on it - the AI will also learn from the photos that are liked on the Internet today.
And these are possibly the pictures of white people with blue eyes.
From this the machines conclude that these people are more important than others, and the more photos they see of people with such attributes, the better they can recognize them and tell them apart.
That is structural racism.
And we won't be able to delete it that easily.
WORLD:
I'm afraid little will change in the men-women disproportion in the IT scene in the next 20 years.
Are you more hopeful?
Ait Si Abbou:
One of my
bosses
always said: Hope is not a strategy.
A great sentence.
Well, of course I have hope, but it is not enough for me, and I will take action and work to ensure that more women are enthusiastic about the IT industry.
I am involved in many initiatives where the aim is to show young girls at “Girls Days”, for example, how great this job is for women in particular.
display
At Deutsche Telekom, I manage the women's network and try to raise awareness there of how important it is for women to be involved in all decisions, and I'm also loud and I use my voice to drive the issue forward.
But we're talking about a social change here.
The fact that there is no equal participation of women and men is a global problem and is related to the patriarchal structure, which is even older than racism.
WORLD:
How is diversity in your team at Telekom?
Ait Si Abbou:
The distribution of women and men in our team is much better than anywhere else.
But that is also due to the fact that, apart from Germany, we work in four Eastern Bloc countries: in Poland, Russia, Slovakia and Hungary.
There it is, as I already mentioned, quite normal for women to study data science, electrical engineering or computer science and also be able to program.
In this respect, our team is a positive exception when it comes to diversity.
WORLD:
Algorithms decide about jobs or lending, they control our preferences based on recommendation lists at Netflix, they look for the right partner at Tinder for us and increase the flight price when the booking is made via an expensive mobile phone.
Do we need an algorithm TÜV?
Ait Si Abbou:
No, we don't need that.
The solution cannot be that we restrict the technology, but that we increase its benefits through controls and quality checks.
It's similar to a bread knife: you can cut bread with it, but you can also kill someone.
Nevertheless, it would not occur to anyone to ban the knife or only allow those with a blunt edge.
WORLD:
The ethically responsible use of AI is an important issue.
But standards and morals also depend on fashions and trends, religions and ideologies.
In China, AI systems are used to control the population and a totalitarian re-education program through a social credit system.
How dangerous is this development and how can it be controlled in a globalized world?
display
Ait Si Abbou:
That is a difficult question.
A universal ethic?
What we have in common are the 30 articles of the UN Declaration of Human Rights.
In theory, all countries adhere to it, but not in practice.
As a user: It is difficult to protect yourself when it comes to services for which there is no alternative, or when the state monitors its citizens with facial recognition software, as in China.
This is why data protection, which is often interpreted as a brake on technical development, is so important.
However, when it comes to machines communicating with other machines, you should trust yourself more.
We don't want to allow autonomous drones with bombs.
Clear.
But we definitely want drones that find people buried after an earthquake.
I would like data protection to be as strict as necessary and at the same time as beneficial as possible.
But because of this fear in Germany of doing something wrong, we prohibit too much as soon as there is a small risk.
WORLD:
Face recognition is a dangerous technology, says Alphabet boss Sundar Pichai.
It must be banned until laws regulate its use.
Clearview AI and other companies pulled billions of photos from social media.
Digital technology provides a wonder arsenal of monitoring and control instruments.
This image analysis AI can guide blind people
That Thomas Panek would one day run for kilometers without a guide dog or companion, he would not have thought possible.
An image analysis AI successfully guides the blind athlete along a lane marking.
Source: WORLD / Laura Fritsch
Ait Si Abbou:
As with the bread knife, it depends on what I use the technology for.
Face recognition on a cell phone or access control in the company is not a problem.
It is a problem, however, when it is used to monitor or examine whether, for example, I have dark circles under the eyes or look tired and will probably be less productive.
You have to weigh up.
But of course I am against banning facial recognition as such.
Sundar Pichai's skepticism about this technology is also due to the fact that there are no laws in the USA, but also in China, for the protection of data and privacy as we know them in Europe.