Although this is theoretical a problem, in fact it isn't one. Lets do the calculation around it. EDTA complexes equimolar amounts of bivalent cations (e.g. Magnesium), so 1mM EDTA will complex 1mM Magnesium ions. Lets assume that the PCR reaction contains 2mM Magnesium.
A typical PCR reaction I do has a total volume of 25ul and contains 1ul DNA template in TE buffer (10mM Tris, 1mM EDTA). The final reaction contains 1/25th mM (or 0,04mM or 40nM) EDTA which can complex 0,04mM Magnesium. So instead of running the reaction with 2mM Magnesium, you do it with 1,96mM, which will not be noticable.
Additionally other errors in setting up (pipettes not accurately calibrated, not setting small amounts into the liquid etc.) easily add up to 5-10% error in total, which has a much higher impact in the Magnesium concentration of the reaction than the EDTA. If you use much higher template concentrations (which is usually not necessary), then you can correct for this, but normally this is not necessary.