A key benefit of symbolic (rule-based) artificial intelligence (AI) is its formal rigor, which comes at the cost of formal modeling effort and computational expensive reasoning. Differently, subsymbolic (datadriven) AI approaches usually outperform rigorous ones in performance but might lead to unsound results. Intersymbolic AI is an emerging field in AI that aims to combine symbolic and subsymbolic AI approaches, exploiting the benefits from both worlds. The scope of the ISoLA 2025 track on "Formal Methods for Intersymbolic AI" is to gather researchers and practitioners from formal methods and (sub)symbolic AI to establish the boundaries of intersymbolic AI and to investigate and clarify the role of formal methods therein.