Comment by NetMageSCW

Actually the heat shield was the exact opposite of normalization of deviance. When the Artemis I heat shield behaved in an unmodeled way, they spent two years analyzing the issue, modified their test system to create all conditions of reentry, came up with a new model that took into account more variables and explained the results seen on Artemis I, then duplicated those results in test to confirm. The condition of the Artemis II heat shield is a sign that they were most likely correct.

I still think they shouldn’t have flown astronauts on Artemis II without an unmanned flight to reduce risk, including other systems like ECLSS as well as the heat shield. But it was the opposite of normalization of deviance.