Context : Recent results by Bensby and collaborators on the ages of microlensed dwarf and subgiant stars in the Galactic bulge have challenged the picture of an exclusively old stellar population , because ages significantly younger than 9 Gyr have been found .

Aims : However , these age estimates have not been independently confirmed with different techniques and theoretical stellar models .
One of the aims of this paper is to verify these results by means of a grid-based method .
We also quantify the systematic biases that might be induced by some assumptions adopted to compute stellar models .
In particular , we explore the impact of increasing the initial helium abundance , neglecting the element microscopic diffusion , and changing the mixing-length calibration in theoretical stellar track computations .

Methods : We adopt the SCEPtER pipeline with a newly computed stellar model grid for metallicities [ Fe/H ] from -2.00 dex to 0.55 dex , and masses in the range [ 0.60 ; 1.60 ] M _ { \sun } from the ZAMS to the helium flash at the red giant branch tip .
By means of Monte Carlo simulations we show for the considered evolutionary phases that our technique provides unbiased age estimates .

Results : Our age results are in good agreement with Bensby and collaborators findings and show 16 stars younger than 5 Gyr and 28 younger than 9 Gyr over a sample of 58 .
The effect of a helium enhancement as large as \Delta Y / \Delta Z = 5 is quite modest , resulting in a mean age increase of metal rich stars of 0.6 Gyr .
Even simultaneously adopting a high helium content and the upper values of age estimates , there is evidence of 4 stars younger than 5 Gyr and 15 younger than 9 Gyr .
For stars younger than 5 Gyr , the use of stellar models computed by neglecting microscopic diffusion or by assuming a super-solar mixing-length value leads to a mean increase in the age estimates of about 0.4 Gyr and 0.5 Gyr respectively .
Even considering the upper values for the age estimates , there are four stars estimated younger than 5 Gyr is in both cases .
Thus , the assessment of a sizeable fraction of young stars among the microlensed sample in the Galactic bulge appears robust .

Conclusions :