I wouldn't be in cognitive neuroscience if I thought our methods held no promise, but there's a vast gap between reconstructing images from visual cortex responses and accurately determining whether a person is lying. Farah et al. [0] recently summarized the best available evidence and concluded as follows (emphasis mine):

"... different policies should be considered for different applications of fMRI-based lie detection. We do not join calls to ban fMRI-based lie detection across the board. Despite the enormous shortcomings of the current evidence ... we suggest that restrictions should be proportional to the outcomes and principles at stake. Risk reduction in dating calls for different standards of certainty and different protections of individual rights than the interrogation of terrorist suspects."

[0] http://www.nature.com/nrn/journal/v15/n2/abs/nrn3665.html