Bias by Design? Gender, Generative AI.

Bias emerges from systems and their use
Generative AI systems are trained on data that already reflects societal biases. At the same time, the humans using these tools bring their own assumptions, stereotypes and blind spots into the process. When biased systems and biased usage meet without reflection, new technology does not create progress but amplifies existing inequalities and outdated thinking.

Default outputs reinforce stereotypes if left unchecked
The lecture showed how image generating tools often default to stereotypical representations, such as portraying experts or professors as older white men. Women, non binary people and People of Color are frequently misrepresented or underrepresented unless explicitly prompted. When these outputs are not questioned, they reinforce stereotypes and risk moving us backwards instead of supporting the progress many brands aim to drive.

Representation gaps can be observed and measured
Bias in generative AI is not just anecdotal. Studies and visual analyses show clear patterns in how People of Color are represented, often simplified, distorted or missing altogether. These patterns persist even after multiple prompt iterations, highlighting how deeply embedded representation bias is within current models.

Critical use is part of AI literacy
Working with generative AI means questioning outputs, understanding the assumptions behind them and resisting the urge to accept results at face value. Without this distance, there is a real risk of blindly reproducing bias at scale, especially in creative and communicative contexts.

Future progress depends on conscious choices
While large scale models dominate the market, the lecture also pointed to research driven and community based initiatives that explore fairer, more inclusive approaches to AI development. Even when specific tools are still emerging, the key takeaway remains clear. The future of AI is shaped not only by what is technically possible but by which systems, values and practices we actively choose to support.

FAZIT
Keeping up with new tools and getting excited about the possibilities generative AI offers in the creative field is important. But it comes with responsibility. Beyond improving our prompting skills, we also need to strengthen our AI literacy and question how these systems are built and used. Otherwise, there is a real risk of reproducing old patterns and reinforcing racist or patriarchal structures, only now with new technology. Progress does not happen by default. It requires awareness, reflection and conscious choices.