Microsoft Innovating Responsibly
Training Series
AI has the potential to greatly benefit our world, but without responsible development, it can also pose significant risks. In collaboration with Microsoft, Authors has developed an educational series for engineers to promote ethical, responsible, and sustainable AI development. By prioritizing responsible AI practices, we strive to contribute to a more equitable and inclusive technological environment.
Visualizing Complex Ideas
Our objective was to tackle technical subjects such as harm mitigation and implicit bias, transforming them into a visual lexicon. The Authors team achieved this by illuminating complex concepts through the development of a visual language, utilizing graphic design and animation techniques. Employing abstractions of triangles and polygons, we symbolized themes such as shattering beliefs or delving deeper for inspection. This approach made large concepts more digestible and challenging topics more accessible.
Company-wide Training
Developing programs for AI training that prevent harms is paramount in shaping responsible AI development practices. This training is essential viewing for any developer at Microsoft involved in AI projects, as it provides a crucial understanding of the underlying rationale behind AI development: to mitigate potential harms. The objective is to cultivate systems that are not only useful but also resistant to generating false conclusions or hallucinations induced by underlying assumptions. By prioritizing this training, developers are empowered to spearhead the creation of AI systems that mitigate biases, thereby ensuring fairness and equity in human-AI interactions.