I think this is rather disputed. The 5g figure appears to have originated with Reuters Singapore based on an Australian study, and it's even pegged as a "conservatives estimate" therein.
On the other hand
researchers from Wageningen University & Research calculated that we ingest about 0.0041 mg microplastics particles a week (less than a grain of salt) and 12.3 mg in a lifetime.
i.e. that most people consume less than the equivalent of a salt crystal of plastics:

According to the abstract of the Dutch study that spurred that large graphic:
Microplastic (1–5000 μm) median intake rates are 553 particles/capita/day (184 ng/capita/day) and 883 particles/capita/day (583 ng/capita/day) for children and adults, respectively. [...] Simulated microplastic concentrations in stool agree with empirical data.
However, the modelling is rather involved, as it entails knowing the distribution of particles by weight etc. So I suspect there might not be agreement on this, in the literature.
According to the Canadian study found by Oddthinking, the largest single contributor appears to be bottled water (it has 22-times the plastic contents of tap water by particle count).
That study says that Americans, on average, drink one 0.5L (well 0.436L) bottle of water per day, with the rest of water intake from other sources. Also
The items considered in this study were determined to have
the following average microplastic concentrations: seafood =
1.48 MPs/g, sugar = 0.44 MPs/g, honey = 0.10 MPs/g, salt =
0.11 MPs/g, alcohol = 32.27 MPs/L, bottled water = 94.37
MPs/L, tap water = 4.23 MPs/L, and air = 9.80 MPs/m3.
As Oddthinking notes, it's not really clear how to translate these into weights. Assuming the particle counts are remotely proportional with with the weight, that makes bottled water the largest contributor.
But a 0.5L plastic water bottle weights like 10 grams, so there would have to fairly large leaching (on the order of 5% by weight) to drink half of the mass of one container even a week if all water drank was from such bottles (14 liters = 28 bottles), and even more leaching would be needed if only one bottle per day was drank (American average), i.e. ~20% leaching.