An independent comparison portal /// a certification body for AI models

compare.ai provides guidance through the jungle of GenAI: as a non-profit organisation, we collect all available analyses, evaluations and critical background research. We then process this information in such a way that users without special technical expertise can explore the diversity of AI models.

The performance of the models is not the only factor that plays a role here; a number of socio-political/economic parameters are also important:

• Transparency and quality of data • Handling of bias • Handling of copyright issues • Environmental sustainability • Social sustainability (subcontractors, data curation in low-wage countries, etc.) • Harmful data filtered out? • Open source or closed off • Are the developers profit/non-profit oriented? • Personality type

Ai-compare is currently still in the beta project phase, so to speak: it is less about results and more about discussion. Which criteria should be weighted how heavily, can all of this be quantified, would a simple ranking be the most user-friendly solution, or should the subjectivity of the weighting/assessment be the focus?