While the result of using a power source with less wattage capability than the Mavic car charger is listed as requiring is uncertain, it seems extremely unlikely that it will work as required. Several outcomes are possible (as Passerby has noted) but as "success" is unlikely to be one of them, the others are largely of academic interest.
I launched into an answer based on the assumption that the battery is essentially a "dumb" LiIon pack. This is not the case.
The Mavic Intelligent Flight Battery contains both a BMS (battery management system) to control cell balance and other charging functions, and also, importantly, a system to control battery functionality such as output power level under extreme temperature conditions. It will also discharge the battery to an optimum point when left unused for extended periods and has a 'complex' interaction with external in flight monitoring and control software. While it MAY be that this battery allows low charge rates and dumb interface, it seems unlikely.
Worst case the risk seems to be the destruction of a $A149 battery** and say a basic $A33 AC charger.
My guesstimate (all care no responsibility) is that charger destruction is unlikely and battery destruction not overly likely :-).
One solution that would work is to charge an intermediate reservoir of adequate voltage with enough capacity to charge the battery at "approved" levels and then apply it to eg a Mavic car charger. The reservoir could be eg a 4S bank of old cheap deadish LiIon cells, or possibly a small lead acid battery or a super capacitor.
A widely available 12V 7Ah "motorcycle battery" would have ample capacity but perhaps not enough voltage across its discharge range - depending on how slavishly the Mavic car charger maintains its 12.7V input voltage.
A reservoir with less Wh (Watt hour) capacity than the target battery needs for a full charge would require multiple charge discharge cycles.
Used 18650 LiIon cells with around 1Ah of capacity are available for under $1 each - and new high quality eg 3Ah + cells for under $10 each. 4 of the latter would probably be able to completely charge a 'flat' MPIFB (Mavic Pro Intelligent flight battery) - and could be recharged by your lower power source either at the same time as being used for charging or between MPIFB charges.
The Intelligent car charger input range is listed as 12.7 - 16V - near ideal for 4S LiIon use. Per cell that's 3.18 - 4V. As long as the cells were not charged above 4V (rather than the normal Vmax of 4.2V). You could try 4 x 4.2 = 16.8V if feeling lucky or adventuresome. But a simple LiIon charger that maintains an Imax charge rate until 16V is reached and then ceases charging would probably be adequate. (Watch cell balance).
Supercapacitor?: At 50 Watts output a supercapacitor rated at say ~= 20V would provide about 2 seconds charge per Farad. Sadly, this is liable to be "very financially unattractive" at current pricing. For example these 12V 5 Farad supercaps cost about $US 60 each. Two in series provides 2.5F at 24V - enough for 2 or 3 seconds charging output to a MPIFB before recycling.
This was my original answer. I'll leave it here (for now at least) as it applies more generally to other systems with somewhat less integration of the battery and system controller.
It is highly likely, but not certain, that the Mavic battery can be charged at rates lower than the car charger's full-rate and probably at arbitrarily lower rates. If so, a "technically good" solution would be to ascertain what the Mavic battery charging requirements are and 'address' the battery directly. The most likely obstacle to this would be some form of custom handshake / interface protocol implemented by Mavic. Such protocols are often explained as ensuring that batteries are always charged 'correctly' by approved equipment. While such a justification has some merit, a more cynical interpretation is that this approach limits the market to those controlled by the manufacturer and allows them to control pricing and capabilities offered.
As the Mavic community is large it is likely that it is publicly known whether a proprietary handshake mechanism exists.
If operating 'blind', one approach would be to examine the output of an 'approved' charging system to see if only battery charging voltages are present. A single charger channel at a minimum requires 2 wires, and may have some means of ramping up under-minimum-voltage batteries, but if presented with a battery in a partially charged state, would be expected to apply a voltage of 4 x (3 <= V <= 4.2V) until the pack reaches a 4 x 4.2V = 16.8 volt "CV" point and then change into CV, decreasing current mode.