There have been some pretty extreme cases reported in the press recently, but two of them particularly caught my eye.

In the first, a man was sentenced to 18 years with a 6 year extended licence, for producing AI images of children and encouraging the rape of a child.
In the other, a man was sentenced to 6 years with a 3 year extended licence, for encouraging a child to commit suicide and possessing illegal images of a child.

What is particularly striking about the two cases is that the difference in sentencing. In the first one, the person was communicating with an undercover police officer. At one point they discussed raping a child, which led to the encouragement charge, although the judge conceded that no child had actually been raped. In the second case, a real child was encouraged to commit suicide, although there is no evidence to suggest that they did go through with it.

To treat encouraging a crime exactly the same as committing the crime yourself is rather odd. It's a kind of joint enterprise thing, but it turns the written word into a deadly weapon, and we all know that people talk about things online which they would never actually do in real life. Also, how is encouraging rape a worse crime than encouraging suicide?

In assessing risk, you would obviously look at likelihood of someone acting on their words, but how does that work for encouragement? These two individuals might well be at risk of saying something very unpleasant in the future, but are they really going to do anything themselves? It seems to me that these are essentially thought crimes. They might lead to an actual crime being committed, but equally they might not.