Rokas Leonavicius offers a fascinating theory about how fanciful, so-called "no touch" martial arts techniques may be born.
I don't think this explains all the no-touch techniques I've seen, such as George Dillman's routine. Perhaps there are more steps or degrees between the conception phase Rokas is describing and the delusions that Dillman and his ilk maintain.
Regardless, this theory suggests two safeguards that can reduce the possibility of such poisonous groupthink.
Bad Guy: Don't Give It Away
Training partners must offer some level of resistance at every stage of training, even initial skill acquisition. That doesn't mean that they must actively attempt to counter, nor that they resist by strength or technique. Rather, as the "bad guy", you must never anticipate how the "good guy" will move you. You must not "give" the good guy the technique, just allow them to "take" it. If they're not doing it right, or could improve, speak to them and describe what/how they should move. But don't reposition your body to be perfect for them. The "bad guy" has to play it a little dumb. Channel this guy's spirit:
When you're playing the "bad guy" consider this question: If your partner was accosted on the way home from the gym today and needed to use this technique to save their life, have you have done everything you could to maximize their chance of success?
Good Guy: Ask For Appropriate Resistance
If you're the "good guy" in a training repetition, and you aren't either initially learning a skill or specifically trying to refine your movements, ask for resistance. Suggest the types and degree of resistance the "bad guy" provides initially to focus your training time. Tell your partner if they're resisting too much. Trying to out-muscle each other is probably not the best use of your time. This may require some negotiation. You may not get it right at first. But you need to establish some mutually understood graduations between "pliable dummy" and "win at all costs." Adopt the nomenclature Craig Douglas uses or Stephen Kesting's, or make up other terminology that appeals to you and your training partner(s). The important thing isn't the specific words you use, but the understanding and trust you have with your training partner.
When you're playing the "good guy" consider the same question: If you will need this technique to save your life later today, have you done everything you could to maximize your own chance of success?
Rokas' hypothesis is also disturbing because it points out one of the critical dangers in learning and practicing "deadly" techniques, those "too deadly for the ring". If you can't employ "randori" or "sparring" to establish the timing, distance, and positional aspects of a technique against a resisting opponent, your chances of executing that technique under the conditions of a real fight are very, very low. I'm tempted to say "zero", but accidents happen.
A specific example is the groin grab/rip/twist/tear/whatever. Almost none of us have ever performed that technique with enough realism to have high confidence in it. What's it like against an adversary in gym shorts vs one in heavy jeans? How much force must you apply? Can you establish your grip and hang on while your opponent is moving around (it seems kinda likely they will)? (My first martial art taught it, too, so I'm not picking on yours!)
If your training partner must pretend to react to your technique, rather than being forced to, Rokas' hypothesis places you one step down the road to no-touch nonsense.