Right now I'd conclude that you can accidentally release things from labs with virtually no consequences. Just make sure that a similar virus exists naturally somewhere on the same continent, don't share any internal records, don't let your staff talk to the press, and you have plausible deniability.