It can also be revealed that the government is working with cutting-edge artificial intelligence and big-data researchers to streamline the movement of passengers, parcels and mail into Australia and better predict incoming threats.

Michael Outram, acting Commissioner of the Australian Border Force, told Fairfax Media that the Sydney case had underscored the need for Australian agencies to co-operate to gather more information about packages before they arrive.

"It was a game-changer for us in the way we look at the border, particularly in a holistic way with partners including OTS [the Office of Transport Security], because it made us realise ??? we cannot just look at the border in terms of the physical border when things arrive here.

"The global supply chain is borderless, frankly. It's a continuum rather than just a physical border at which our people have been operating traditionally.

"So we very quickly had to partner with OTS and look at where we could put screening technology offshore. But to do that, we need to understand what happens at the various airports and choke points in the supply chain that, say, DHL, UPS, TNT or FedEx use and what are the arrangements at the various airports at which these parcels are being embarked onto aeroplanes."

In early August, police announced that as part of an Islamic State plot to bring down an international airliner, terrorists in Syria had allegedly sent military-grade explosives to comrades in Sydney via air freight from Turkey. The discovery left seasoned national security observers amazed.

Mr Outram said he had already met with the major carriers about this and all had expressed a strong willingness to work with the government to ensure protection of the community.

It is unclear whether the government is looking at stationing Australian customs or transport security personnel at overseas transport hubs.

While Border Force is concerned with whether an item coming into Australia is prohibited, the OTS regulates security for inbound and outbound planes and ships. They will soon be side by side under the new Department of Home Affairs.

Mr Outram said Border Force had been "push[ing] the envelope further out in the supply chain" by gathering as much early data as possible on incoming passengers and packages "rather than physically moving people offshore".

He added: "But certainly we'll have to continuously reconsider and rethink our approach to supply chain security, particularly with OTS and [Border Force] moving into Home Affairs ??? and how we can work together and collaborate and share capability and information and data."

Transport companies operate according to complex logistics models that shift constantly. Each carrier has different routes on different days depending on the volumes coming from countries and how these are consolidated at international hubs.

In theory, everything that goes on an international flight is screened before it is embarked but standards and practices vary around the world.

Some screening is done by local customs agencies, some by freight companies, some by airlines and some by airport security operators.

Customs in Australia cannot check everything - the volume is too large. Last year 42 million air cargo consignments came into Australia. The number of inbound people and packages is growing at double-digit percentages each year.

"The magnitude of those increases are such that you just cannot deal with them with more officers," Maria Fernandez, the Department of Immigration and Border Protection's head of intelligence, told the recent Informa national security conference.

Authorities already use intelligence-based risk assessment to prioritise people and items for scrutiny. The department's Strategic Border Command and the Border Intelligence Fusion Centre crunch and analyse data to produce profiles and alerts.

But this would be dramatically increased by investment in cutting-edge machine-learning and big-data analytics, Ms Fernandez told the conference.

The department is working with the CSIRO Data61 innovation centre and Data to Decisions CRC - a co-operative research centre involving government and industry - on introducing this.

The more information they can get, and the earlier, about who and what is bound for Australia, the better they can create risk profiles.

But Sanjay Mazumdar, chief executive of Data to Decisions CRC, said the sheer volume of data from a wide range of sources - not all of it nicely structured such as cargo consignment spreadsheets - meant authorities risked "drowning in data", a challenge faced by all security agencies today.

"The scale is growing too rapidly. Manually trying to do it is impossible, and so what they're looking for is more sophisticated techniques really to filter down all that data and present the investigator or the customs officer with 1 or 2 per cent of risky cases rather than the 98 per cent of clean cases."

The kind of data that is useful obviously includes things like the place of origin of the package and any sender and addressee details. But it can include all sorts of other anomalies, including "dwell time in ports, inconsistencies in routing and other information taken from consignment documents", according to Warren Bradey of the CSIRO's Data61.

"For example, data analytics could quickly learn if a container has been held in a port for an unusual length of time," he said.

"Packages have a great deal of data associated with them beyond just the sender and receiver note. Machine learning finds patterns between the enormous amount of data associated with a package and 'unusual' events."

Mr Outram said all of this was compared with - or "washed against" in security jargon - intelligence information.

"If we get an early heads-up from the intelligence agencies around the world, including our own here in Australia [that] - ISIL or a criminal group are adopting a new method or are intending to do something different or new - we can adapt our practices and processes and profiles and alerts, to be on the lookout for it without grinding the border to a halt."

In the case of people, who have unique biometric data such as faces, the next generation of immigration "smart gates" at airports would allow people to walk through without stopping, Ms Fernandez said.

"By the time someone gets off the plane, with face-on-the-move technology, which our new-generation gates will allow, the gate can be open and someone can walk up to the gate and just with their face, the gate will be able to go - 'I know you. I've got your passport information and just with your face as your passport, we know who you are. You hit no alerts. You've hit no profiles. You've hit no predictive models.'

"The gate will remain open and you will just walk through a seamless, open border. Only if you hit on something will the gate close."

As well as crunching much more data quickly, learning computers can also find patterns or anomalies that a human wouldn't have thought to look for.

"Human-designed rules are based totally on the past experience of the agency in having seen the origin and pattern of where suspicious items come from," Mr Bradey said. "Whilst that is an invaluable source of information, it does not allow real-time learning of emerging trends or anomalous activity detection for new occurrences."

Dr Mazumdar said computers were good at "picking up things that might not necessarily be perceptible to the human".

"That one pattern that might be in the noise - a computer if trained well enough may be able to pick that up and present that to the analyst.

"You might actually surface up stuff they wouldn't necessarily find because they haven't correlated seemingly uncorrelatable data."

That may mean drawing in disparate "pattern of life" data, he said. As an example in immigration, a person who repeatedly enters the country with several female passengers and then leaves the country alone may be engaging in sex-trafficking.

"Training the computers to look for those patterns amongst the data available to them is something a computer can do that a human can't do at scale," Dr Mazumdar said.

Jacinta Carroll, a former security official now with the Australian National University, said understanding where the risks were allowed authorities to shift resources around as the threat landscape changed.

"It means being able to ratchet them up and focus on particular things when you need to," she said.

She stressed that technology would not solve everything. Terrorists and criminals had shown they could circumvent or even exploit new technologies.