Consciously, sure, but with some studies saying judges tend to rule differently on similar cases based on whether it's before or after lunchtime, do you really want to introduce more subliminal variables like decorations into the legal process?
The "hungry judge" finding did not replicate [0], and the original paper had serious methodological flaws, namely, that prisoners with parole hearings later in the day were more likely to have either overworked "shared counsel" (the Israeli equivalent of public defenders) or no counsel at all [1]. You should consider the hungry judge effect another debunked victim of the replication crisis.
The influence of cartoon dragons probably don't come close the influence of judges and lawyers talking to each other about what schools they went to, what country clubs they belong to, etc.
The breech of formality isn't being turned into a big deal because it might bias case outcomes; in that regard it's a rounding error washed out by innumerable more substantial sources of bias. It's being made into a big deal because formality is the wall between outcomes of cases and feelings of personal culpability for the people who are involved in that process. All of the formality and decorum make it easier for judges and lawyers to emotionally distance themselves from, very often, ruining peoples' lives.
Too many "rounding errors" adding up seems like a reasonable justification for not allowing even seemingly small sources of additional inconsistency. Slippery slope and all that.
Also, just because the other potential sources of bias you brought up exist doesn't mean new ones should be let into the process. I wouldn't be against solutions to remove the ones you mentioned. But I don't think you'd be entirely convinced that just allowing cartoon dragons would decrease bias by making people more empathetic.
Do you know for a fact that those rounding errors add up, or do they cancel out? I feel like all this bikeshedding is just helping to draw the attention off the more significant and well known biases.
Do you know for a fact that the rounding errors cancel out perfectly? Probably not.
Trying to make this about "more significant and well known biases" when all I'm arguing for is not introducing additional biases is a logical fallacy. And allowing cartoon watermarks in court documents would not help with the "more significant and well known biases" that you're more concerned about, anyway, so I have to wonder why you feel strongly enough that this should be allowed that you'd write off opposition as "bikeshedding."
