Lithium plating affects four key characteristics of cells: life-time, low-temperature performance, how fast they can be charged, and safety.
The insertion potential of Lithium ions in graphite is in the range of 200 to 65 mV vs. Li/Li+ [metal deposition]. For potentials below Li/Li+, Lithium deposition on the graphite anodes becomes thermodynamically allowed.
I think it's more correct to say that if the potential of Lithium ion insertion (intercalation) is higher than the potential of Lithium metal plating, the latter can still happen (according to the Butler-Volmer equation), but the rate will be negligible.
In thermodynamic equilibrium, the cell voltage can be determined by the Nernst equation, however, a deviation is observed experimentally which is the polarisation. The main causes for polarisation are Ohmic drop, charge transfer overpotential, and diffusion overpotential.
I think there is some confusing terminology used here. What the authors call "polarisation" Wikipedia calls "overpotential", whereas "polarisation" is just one of the causes of the overpotential (polarisation is the effect of Cell voltage hysteresis).
The potential of Lithium intercalation decreases, and, hence, the risk of Lithium plating increases when charging either 1) at low temperature, 2) with strong current, 3) at high State-of-Charge level.