It depends of the finer technicalities, for example:
Connecting to an unsecured wifi network and using it for internet access is OK.
This is because your device asked the AP for permission and the AP responded giving permission. So as you asked for permission and got it you are in the clear. Open wifi networks still have authentication, it's just that the routers policy is to grant access to anyone who asks for it.
Figuring out that an SQL server is listening on an open port and sending SQL queries to it is NOT OK.
Because there was no authentication process at all, so you didn't have permission and it is therefore a crime.
I'd better put on my 'not a lawyer' hat first, but I think this misses something, because it would make any social engineering attack exempt. The word the CMA repeatedly uses is "unauthorised", "authorised" appears a couple of times. Interpreting a computer's automatic response as granting authorisation is probably wrong, interpretation notes in 17.8 : "An act done in relation to a computer is unauthorised if the person doing the act (or causing it to be done)—
(a)is not himself a person who has responsibility for the computer and is entitled to determine whether the act may be done; and
(b)does not have consent to the act from any such person.""
Authorisation comes from a person, we might use the same words for processes a computer carries out, but they're not the same. An open wifi is authorised to connect to because its owner has chosen to make it open. If it's been left open by accident there might be an argument for that being unauthorised access. And this is where I'm really not a lawyer, how much is thinking you had been authorised a reasonable defence? And implausibly claiming you thought you were authorised when you weren't, such as stealing a neighbour's wifi may not go down well. However prosecutions under CMA are rare, so I don't really see the police getting involved in that one (inconvenience a company on the FTSE and the story might be different).
Ihttp://usir.salford.ac.uk/15815/7/MacEwan_Crim_LR.pdf makes interesting reading (and serves as a reminder that lawyers aren't always in agreement about interpretation, see for example the section on DPP v Bignell)