First, I recently retired from a major auto mfgr after 35 years all in airbag safety restraints. I wrote most of the specification for the inflator portion of the airbag specification. I have designed inflators, built plants that make them, managed release groups that put them in cars and tested thousands of units. So in short, I know my stuff.
There is just too much here to comment on, so I will just add a few general principles;
First the company (now bankrupt) with the great airbag recall was Takata, not Tanaka. The issue is/was degradation of the solid propellant over time due to moisture intrusion and temperature swings during the life of the car. The issue is specific to one of the Takata propellant formulations.
Now for the good stuff:
Nothing can protect an occupant if the occupant compartment does not stay intact. Crush is a good thing as long if it happens outside the occupant compartment since it spreads out the forces and reduces the G load. Try this with an egg. Drop it on concrete, then add a washcloth to the concrete. It doesn't take much distance to lower the G load.
Injuries do occur to the brain with little to no external injury. Google 'closed head injury' if you do not believe me. This generally involves abrupt twisting motion and can couple to lateral motion as well. The brain is 'floating' in the skull and can tear critical attach points as well as internal structures with no external injuries.
Seatbelts are your first, and in low G crashes only defense. The issue with belts (3 point) is the 2" width and elongation. Elongation of belt webbing is typically 10-15% under typical restraint loads but increases with load. With a retractable belt elongation has to be considered for the entire belt length as well as the spool out before the retractor locks. The width issue is spreading the load out on the body. Wider is always better. Multiple straps are always better so long as the belt is positioned correctly. Keeping the lap portion of a 3,4 or more point harness is critical as the belt needs to restrain on the pelvis rather than soft belly. The shoulder straps must rest on the shoulder to restrain properly. The belt loads in even mild crashes can easily exceed 2500lbs, so attach points must be secure. DO NOT ever attach a seat belt to a seat back unless the seat back is a structural unit the can withstand the 2500+ pound forward load.
Inflators for any airbag or belt are either solid propellant, stored gas with a pyrotechnic heating element, or straight stored gas. All are deployed based on a G sensor oriented in the direction of the intended protection. i.e. will command deployment with a longitudinal G load, but not vertical. The G load signal is 'filtered' electronically or mechanically. Inadvertent deployments do happen when conditions outside the signal filtering happen or there is damage to the system. Ruptured inflators have been extremely rare (except in Takata's case) I have personally investigated and determined cause on most cases involving inadvertent deployment or inflator rupture with my previous employer over the last 25 years. The risk is much lower than crashing your airplane, but deployment injuries do happen.
there is so much more, but enough for now. To the original question, having the additional protection of a seatbelt airbag is a trade off. Comfort during wearing, cost and added weight against the risk of a crash for which the belts were designed to provide added protection.
I will answer any questions relating to safety restraint systems (belt or airbag) but please PM me to start a new thread so I don't hijack this one.