Archive

Archive

Blog.Telekom

Elisabeth Giesemann

0 Comments

"What's wrong with our data?" Women in AI at the Startupnight

Machines learn from people and adopt their patterns of thinking. Yet, since women are underrepresented in the field of artificial intelligence (AI) development, the machines are only programmed by part of the population. The participants of the panel on "Women in AI" a the Startupnight 2018 discussed this problem.

Women in AI at the Startupnight

Artificial intelligence will change our society from the ground up. Therefore, the participants of the panel want a diverse programming of the AI.

"What is wrong with our data?" Asked Kenza Ait Si Abbou Lyadini, Senior Manager Robotics and Artificial Intelligence at Deutsche Telekom IT. The panel's moderator is also on the board of Deutsche Telekom AG's umbrella organization for women's networks. She notes that although they should be objective, the results of AI often exacerbate prejudices and biases.

As the past has shown, a lack of diversity in the development of AI systems leads to discrimination, for example in recruiting or facial recognition software. Search engines have shown well-paid job offers only to men.

"Our technology reflects society, which is why artificial intelligence is just as influenced by prejudices as humanity is," explains Dr. Susann Wegner, Chief Data Officer at Deutsche Telekom. She wants to further establish the data business within Deutsche Telekom in order to scale it across the Group. Therefore, the advancement of women in this area is important for her both personally and professionally.

Whoever programs the AI is therefore decisive for the result. "You can't expect morality from machines, they have no awareness in this sense. Machine learning is first and foremost about making processes happen faster and replacing unpleasant tasks," stresses Claudia Pohlink, Head of AI at T-Labs. For this reason, neutral data cleansing and processing is essential for a more diverse design of AI.

"Most data scientists are male and of course this has an impact on the technology. To promote women in the field of AI, Karolina Attspondia founded the W-Lounge, a platform that enables women to network in technology and business.

Elena Poughia, the founder of the Data Natives conference, also sees networks and female role models as important drivers for positive change: "You need more women at conferences and on stage. Women need to see that they can do these things. There are many different ways to contribute to the development of AI and they can be part of it".

"It's also sad that the programmers don't realize anything is going wrong until they've created a sexist or racist monster. So far, we have many of the men in hoodies who program these technologies and we want to change that." says moderator Ait Si Abbou Lyadini. The panel therefore also addressed the question of how more women can be persuaded to develop AI. Claudia Pohlink explained that education of young women and girls is especially important. However, they should not be pushed in one direction, but the freedom to experiment is essential for development, so that they actively participate in the design and program a diverse AI that corresponds to our reality.

"You can do everything professionally and be active in AI. AI is so diverse and we also need women with backgrounds in business administration, communication or medicine who contribute to the developments," explains Ait Si Abbou Lyadini. Poughia also adds: "A straightforward curriculum vitae is becoming less and less important. If the machines are faster and more effective, creativity will be crucial in the future.“

The "Women in AI" now want to create a space for experimentation. On November 15 and 16, 2018, they will organize an AI hackathon in which women from all professional backgrounds are invited to experiment together and develop AI-based solutions. The event takes place at Hubraum, Winterfeldtstraße 21, 10781 in Berlin, Germany, and is sponsored by Deutsche Telekom.

Roboter in lauschender Haltung

Does Artificial Intelligence have feelings?

Almost no other aspect of digitization instills as much fear as Artificial Intelligence (AI). We examine this question and have asked two experts.

FAQ