We're in like year 2 or 3 of LLMs being serious tools, still with many disadvantages over humans. There's plenty of time to figure things out, we don't need to experiment on children right now.