I posted previously that Tosafos holds that a goy is not chashud to cheat without any benefit, even if it won't be detrimental to him either. http://hearos.blogspot.com/2010/09/avoda-zara-31b-relying-on-kashrus-seals.html
It is difficult to figure out Rashi's opinion on this issue. The gemara 34b that says goyim aren't suspected of adding wine to murayus because קסתא דמורייס בלומא קיסתא דחמרא בד' לומי implies that only when it cost more to add wine will they be deterred from doing so, but for no gain they may cheat. Rashi seems to assume like this because when the gemara says that the first and second making of the murayos "doesn't need" wine to enhance the flavor, rashi writes - איידי דנפיש שומניה פעם ראשון ושני מקלקל להו חמרא. Why would Rashi have to say that the wine would "ruin it"? Rashi seems to hold that even if the goy has no reason to add wine, he may do it anyway.
Rashi in our sugya also seems to follow this approach. The gemara says that there is no reason for a goy to swap the kosher bread for the non-kosher bread. Rashi writes that since he has no benefit in swapping it, one seal is sufficient - כיון דאיכא חותם אחד, לא טרח ומזייף במידי דלית ליה רווחא. Since he has nothing to gain, why would we require any seal at all? This question is raised by many achronim (mateh yehonasan [r. yonasan eibcschitz] in y.d. 118:10, gilyon maharsha, darchei teshuva). The achronim suggest that although he has no monetary gain in swapping the breads, he may be motivated to do so because he is hungry now and later replaces it with treif bread so one seal is enough to avoid the practical motivation. However, rashi may go li'shitaso that to permit it without a seal, it is not sufficient that the goy has no benefit, rather we require that the goy would actually suffer a loss. The Ritva quotes the ramban that Shmuel in our sugya argues on Rav and permits the bread without any seal at all - this would be consistent with Tosafos 12a that whenever there is no motivation to swap we don't have to be concerned.
See Rashi 64b d.h. aval, where he seems to hold that if the ger toshav has no benefit, we don't have to be concerned for a swap. This would contradict the above approach, but perhaps we are more lenient for a ger toshav than for a goy.