How does militants communicate securely?

How? They are not using the big obvious systems at all. There are lots of entrepreneurs who set up systems for libertarians to use and the terrorists quickly identify these. Systems such as f.e. SureSpot offers consumers an easy way to use encryption. The availability of encrypted systems makes the security agencies crackdown "absolutely pointless". They are all now using the OTR (Off the Record) protocol, which offers end-to-end encryption. Even if you managed to stop companies providing OTR, there are plenty of free-add-ons available. On jihadi bulletin boards, there are links to online encryption tools that people can download. Any jihadi worth his salt to going to know to find a safe way to communicate. Most don't tend to use iMessage or WhatsApp.

Why are the law enforcement agencies so focused on how Apple and Google use encryption?

The big tech firms are low-hanging fruit and it is a case of having to start somewhere. If you want to boil the ocean, you have to do it one cup at a time. There has been a degree of misinformation about how those behind the Paris attacks communicated with each other. It was more than likely they had used WhatsApp, but there is no evidence they did. Meanwhile, a comment from Belgian Federal Interior Minister Jan Jambon that they had used a PlayStation 4 to communicate was widely reported, but it later emerged suggestions the console had been found at the apartment of one of the attackers were false. It was also reported the PlayStation 4 used end-to-end encryption, but it does not.

How are the intelligence services monitoring extremist activity if they cannot read messages?

Encrypted products still reveal metadata - who talked to whom and for how long, and this has played a crucial role in the aftermath of events in Paris. The arrests that are going on now have come from a trawl through metadata. The first person is identified, and then the security services use what is called link analysis to build a picture of who they have been speaking to. The British intelligence service is particularly good at link analysis. And the Draft Investigatory, currently going through Parliament, would require internet service providers to hold on to this metadata for 12 months. But increasingly the security services are turning to hacking - placing malware on suspects' computers to find out what they are doing in real time. If they can hack into a computer or smartphone, then they can find the key to decrypt their messages. There was a militants operative in Syria who was sucked into a Skype scam, persuaded to talk to what he thought was a pretty girl, but she was a cover to inject malware onto his machine. That was able to head off a couple of attacks.

The government raised the possibility militants might launch a cyber-attack in the EU/UK. Should we be worried?

For the past 20 years, there have been concerns a terrorist network could cause real disruption by attacking critical infrastructure such as air traffic control or hospitals, and Chancellor has just announced he is going to double funding to fight cybercrime to £1.9milliard a year by 2020.
He said militants were trying to develop the ability to launch deadly cyber-attacks, but his comments were "vague" rather than being about specific threats. There are more power outages caused by squirrels than by cyber-terrorists. However, the Stuxnet malicious worm believed to have targeted Iran's nuclear programme in 2009 suggests such attacks are possible. These militant groups are getting much more sophisticated, and you only have to look at how they use social media to see that they are very tech-savvy. And much of the government's £1.9milliard budget will go on finding highly skilled people to work out what militant groups might do next.