AI-Democratization: From Data-first to Human-first AI

In recent years, there have been significant advances in artificial intelligence (AI) research. However, not everyone can benefit from these advancements because it requires a significant amount of expertise to apply them. The concept of democratizing AI, known as AI democratization, aims to make AI accessible to everyone. One exciting development in this field is the shift from a model-first to a data-first approach. Previously, the focus was on building and training complex models, which required a high level of expertise. However, new high-level frameworks have drastically simplified the process, making it possible to create machine learning (ML) models based on data descriptions automatically. This has paved the way for data-first AI, where the focus is on the data engineering rather than model creation. The next logical step in this progression is a human-first approach to AI. By simplifying the process of creating ML models even further and leveraging the insights gained from data-first AI, we can make AI accessible to an even broader range of users. We underpin our theoretical considerations with the development of the Open space for Machine Learning (Os4ML) platform and summarize our achievements so far in this endeavor. Our approach has the potential to significantly lower the barrier to entry for using AI and to accelerate AI adoption across a wide range of industries and domains.

34th Central European Conference on Information and Intelligent Systems