Diversity in Tech Awards 2024 : AI & Emerging Tech, What to do about the biases?

Ai

Join experts exploring how to combat bias in AI development through intentional DEI practices, diverse teams, and systemic changes for more equitable tech innovation.

Key takeaways
  • AI systems can perpetuate and amplify existing biases if trained on biased historical data or developed without diverse perspectives

  • Companies need intentional DEI practices throughout the AI development lifecycle, including diverse data curation, testing for biases, and representative development teams

  • Affinity bias leads people to favor those similar to themselves - in hiring, recognition, and opportunities. Organizations need structured processes to counter this

  • The burden of education about bias and inclusion often falls heavily on underrepresented minorities - this needs to shift to become everyone’s responsibility

  • Open source collaboration and industry-wide partnerships are crucial for developing more inclusive AI systems rather than siloed efforts

  • Language and cultural biases in AI systems can disadvantage non-dominant groups through reduced functionality or discrimination

  • Resource groups and employee networks play an important role in retention and creating inclusive workplace cultures

  • Organizations should implement concrete practices like diverse slate requirements for hiring and feature blinding in AI development

  • Personal strategies for dealing with bias include focusing on what you can control, building connections, and taking up space authentically

  • Leadership needs to move beyond diversity metrics to create truly equitable and inclusive environments through systemic changes and accountability