How LLMs influence the gender bias portrayed in the media? (Lightning Talk ⚡️)

AI development is unlikely to stop in the near future, however, even at this stages many instances of sexism can be seen in the products created by the generative AI models. The talk will explore how potentially discriminating data is present both in input and the output of neural networks: from databases to the structure of the layers themselves (for example - working on datasets which are formed by scraping data off websites, such as Reddit or Wikipedia, which have a predominately male demographic). Key terms will be defined and AI developments will be placed in a wider context, which also included new retouching techniques. Finally, possible answers to this problem will be offered, keeping in mind that many solutions - such as increasing computational power or the dataset size - result in the apophenia phenomenon; increasing societal visual literacy will be presented as a possible way forwards.