The calculus isn't that much different if you include non-fatal incidents. There are roughly 200 times more non-fatal accidents than fatal in the US. So this would put the societal spend at at say $67000-$125000 per incident prevented, if the ratios hold equal for all types of incidents. This is still a pretty significant cost, but I suppose it's closer to break even.
The quick search I did says that the "average" car accident has a cost of $10k, and "severe" ones can be more than $100k. Surprisingly (or not) it's hard to find new data; I found links for a 2014 NHTSA study that says $840B total economic cost (direct plus indirect) for that year's accidents, while the National Safety Council says $440B but I couldn't find the paper to get a year or total number of incidents.
I did find a direct NHTSA paper from 2010, that says the "total economic cost" of what looks like 13.6 million total incidents at $242B, or about $18k per incident, or a cost "including quality of life" at $836B, which would be $62k/incident. Interestingly critically injured survivors cost society almost as much as fatalities ($1.0M vs $1.4M)
The values I was using for the cost of TPMS and backup cameras were from around 2010 because that's when they first became really mainstream, so society still seems to be overpaying by a bit, although not as much as I had initially thought.
Personally I think it would be better to have better training than using engineering and manufacturing resources for these technologies as they are mostly only indirectly related to efficient travel. Society doesn't have infinite resources, after all...