I would abandon the word "force" as soon as getting to relativistic physics, and the electromagnetic field is utmost relativistic. I'd rather use the word "interaction".
The discovery of the notion of fields has an experimental basis and some strong hindsight from "natural philosophy" of the 19th century. I think it's pretty fair to say that the discoverer of the field concept is Faraday, questioning the Newtonian idea of "forces" as "actions at a distances" just on heuristic grounds but based in an enormous empirical knowledge from accurate experimental studies of electromagnetic phenomena. Then Maxwell took these ideas up and discovered his famous set of equations describing these phenomena accurately. The trouble was that his theory is not Galilei invariant. The solution of the puzzle famously is the (hitherto final) formulation of special relativity by Einstein (based however on previous work by Lorentz, FitzGerald, Poincare, and others, who fell just short to get the right physical interpretation). Since the special-relativistic space-time structure implies a causality concept that forbids "action-at-a-distance forces" in the Newtonian sense, the most simple descriptions occur using the field concept and substituting action-at-a-distance forces with local-interaction-forces.
The picture is that instead of the "action-at-a-distance-force" acting between electric charges (instantaneous Coulomb force), an electric charge distribution, somewhere localized, implies the existence of an electromagnetic field around it, which is a "fundamental dynamical entity" as the "charged matter" itself. Maxwell's equations then describe the very laws connecting the localized charge distributions with the electromagnetic field. A symmetry analysis based on Noether's theorem shows that the system of charge-current distributions and the electromagnetic field has clearly defined contributions to energy and momentum (as well as stress) by the matter and the fields, and the corresponding energy-momentum-conservation laws translate into local conservation laws in terms of corresponding Noether currents (energy and momentum densities and its currents), which provide the local "forces" acting on the charge-current distributions (as well as the back reaction to the field). These "forces" in this way are made local, i.e., the "force" acting on charges and currents is "caused" by the presence of the field, and energy-momentum conservation thus becomes local and thus consistent with the constraints on causality implied by the relativistic space-time description. That makes it well justified to abandon the notion of "forces", reserving this word for the "action-at-a-distance interactions" in Newtonian physics and talk about "local interactions" in the relativistic case.
There were attempts to get rid of the fields again and finding forms of action-at-a-distance forces compatible with relativity. A famous example is the Wheeler-Feynman absorber theory using half-retarded-half-advanced actions between charges eliminating the em. field. This non-local classical theory can be indeed interpreted in the way making it consistent with causality, but as Pauli famously predicted, Wheeler and Feynman could not produce the promised "quantization" of their absorber theory.
Finally it turned out that the up to today only working quantized theory of electromagnetic interactions is quantum electrodynamics, where all entities, "radiation" and "matter", are described by quantized fields (e.g., in the minimal QED model, a quantized Dirac field and the electromagnetic field with an interaction term provided by the minimal coupling ansatz based on the idea of conjecturing a Dyson-renormalizable quantum field theory).
Thus, as it seems today, Faraday's intuition that Newtons "actions at a distance" are fictitious and to be substituted by "local interactions" mediated by fields has prevailed in a surprising day, providing the most accurate theory ever, the Standard Model of elementary particles.
Unfortunately, it's incomplete since no working theory to also implement gravitational interactions. There's of course General Relativity giving an accurate description of gravity on the classical level in terms of a dynamical spacetime structure. Unfortunately all attempts to quantize this theory, most famously string theory and loop-quantum gravity, have not yet lead to a truly satisfactory description. Unfortunately there's no clue from observations, how a future successful quantum theory of gravity might look like :-(.