FOSS4G 2022 general tracks

Artificial Intelligence (AI) has made an impact in almost every field and has become an incredibly powerful technology. However, while some players in big tech and academia have access to a highly skilful workforce that can create sophisticated AI solutions, many companies, governments, public organisations, and other societal stakeholders are lacking sufficient AI expertise and know-how.

Many AI and Machine Learning (ML) frameworks aim to simplify and democratise AI development, even if typically focusing on those with software engineering skills. Some of these solutions, those that provide no-code tools, get the closest to the ideal of enabling "any person without prior training”. Collectively, these could represent a major breakthrough, as it has been proven time and time again that many businesses and organisations still struggle to implement AI to its full potential and scale.

Visual, often drag-and-drop, no-code AI tools can make AI less intimidating and more comprehensible to non-technical profiles and those who lack the time and resources to build such systems from the ground up. No-code AI frameworks are expected to require minimum technical knowledge to develop practical AI solutions at scale. This is an emerging field, like was previously the case with no-code web development, starting with Dreamweaver and MS Frontpage, the first WYSIWYG (what you see is what you get) solutions, both launched in 1997.

The European Commission is supporting the establishment of an AI-on-demand platform that will provide easy and simple access to AI tools that are made in Europe and are ‘trustworthy’. The platform will gather all the AI resources (algorithms and tools), and make them available to the potential users, businesses, and public administration, with the necessary services to facilitate their integration.

Within the context of the EU’s commitment to trustworthy AI, we are exploring the landscape of AI solutions for non-experts, including no-code, low-code, AutoML and similar approaches, and evaluating them in an experimental setting via prototyping. Our findings are expected to inform the development of relevant European digital initiatives. For instance, we will consider how these emerging AI solutions for non-experts could be integrated and used in the context of digital infrastructures such as EuroGEOSS or the European Data Spaces.

Further democratization of AI will happen when domain experts without prior AI expertise are enabled to tap into high-quality data to solve complex problems on their own, with technical details being abstracted away, such as algorithm selection, model training, software frameworks, hardware dependencies, and platform aspects. We expect that open-source AI technologies for non-experts represent a step towards such a future.

During the conference, we will present the initial results of our geoAI activity, including a high-level architecture and a stack expected to be composed of various open-source software tools and open standards for digital infrastructures, data engineering and AI development. Our intention is to gather feedback from the audience and establish possible future collaborations in this space.