Conventional wisdom says "no", lack of new predictions being the main criticism. Johansson and Matsubara review string theory from various empiricist perspectives, and the best they can say (for string theory) is that it is not the right time to drop it. Yet. They add that it still dominates physics only because "those involved have strong realist convictions".
But many conventional criticisms give me an impression of missing the point. Does one really need to believe in reality made of math, like Tegmark, to embrace unification, with or without new predictions? At the time of Galileo experimental results were at a premium, and theoretical speculations in the absence of empirical constraints were plenty. Today the situation is nearly reversed, experimental methodologies are plenty and routine, we have far more data than we can process, let alone organize and put to use. Experiments in modern chemistry, for example, are little more than analog substitutes for quantum mechanical computations that we are unable to perform efficiently, no one seeks to falsify QM there. The vast empirical input condensed in the Standard Model (SM) and General Relativity (GR) is far more constraining on any unifying theory than any new test can possibly be.
So constraining that so far not a single mathematically coherent (not just heuristic) string theory that reproduces SM and GR in appropriate limits has been proposed. And even the Standard Model itself is riddled with inconsistencies and ad hoc tricks, new tests aren't going to fix that either. Predicting new phenomena is nice, but that was never string theory's main motivation, the incompatibility of SM and GR was. Smolin, one of string theory's chief critics, admits that it unifies SM, and unlike it at least has resources to incorporate gravity too. Moreover, common expectation is that phenomena unexplainable by SM and GR do not manifest below the so-called "Planck energy", which is so huge that it will remain inaccessible for decades anyway, for practical reasons.
The same logic of scientific method that emphasizes empirical testing at early stages of a science discounts it at more advanced stages, when theoretical unification of already available data becomes more important. Since unifying theories maintain contact with reality through already tested lower level theories their independent testing may no longer be a priority, shouldn't they be judged on fulfilling their theoretical goals instead.
Question: Realists may like string theory for revealing truths about hidden reality they believe in, but why shouldn't empiricists love it as a vehicle of unification even if it makes no new testable predictions (which is unlikely)?