Autonomous vehicle (AV) frameworks such as safety critical expert systems and machine learning operate within a designated operational design domain. The former requires careful consideration of all possible action scenario pairs which might cause conflicted behaviour of an autonomous system. The latter struggles with previously unencountered scenarios which may result in suboptimal or critical behaviour. Further complexity is introduced when considering multi-agent coordination and adaptation. To tackle the above challenge, we investigate the use of the Social Force Model (SFM) paradigm to AVs both individually and for group control. The SFM was initially developed for crowd modelling and has been used to demonstrate the interaction between groups and social behaviours such as platooning and lane forming. The social force interaction properties of the model have been recently adapted to consider both vehicle and pedestrian interaction in shared spaces. The research we propose aims to apply and generate such phenomena relating to AV and non-AV agent interactions in traffic environments and thus enable the AV to be more agile and adaptable. Furthermore, we propose the inclusion of a proactive intent-based force in the SFM to: improve the ability of intention prediction models to scale to multiple actors within an environment, provide a base motion policy for all agents in the system to prevent plan-complete terminal states, optimise model parameters to improve outcomes of local and high-level system objectives and improve the safety of lane merging, allowing for the passage of emergency vehicles or optimising AV travel through social grouping and platooning.