The impact of AI & HIoT related threats and recommended approaches

An interview with Richard Staynings, Chief Security Strategist at Cylera at the HIMSS AsiaPac 19 conference in Bangkok, Thailand.


The following article first appeared in the Show Daily of the HIMSS AsiaPac19 conference


Currently leading healthcare security strategy at Cylera, a biomedical HIoT security startup, Richard Staynings has more than two decades of experience in both cybersecurity leadership and client consulting in healthcare. Last year, he served on the Committee of Inquiry into the SingHealth breach in Singapore as an expert witness. He recently spoke to Healthcare IT News on some of the current developments in healthcare cybersecurity.


Q. Artificial Intelligence (AI) applications in healthcare are all the rage now, and so are cybersecurity threats, given the frequency and intensity of healthcare-related incidents. In particular, some of the cyber-attacks have become more sophisticated through the use of AI to get past cyber defenses. On the medical devices front, AI is also being used to constantly manage and secure the rising number of healthcare IoT devices as they connect and disconnect from hospital networks. How do you think the application of AI in healthcare cybersecurity will be like in the next few years?

A. Healthcare is widely considered to be an easy and soft target because “who in their right mind would attack the weak and defenseless?” …. or so the thought goes! The fact is that healthcare presents a rich target for cyber criminals because of the value of the data hosted and processed. When you couple that with a chronic historic under-investment in the development of capable cybersecurity teams and tools across healthcare, you can see why perpetrators are so keen to break in. But it’s no longer the theft of medical records, or PII that concerns me, it’s the wholesale theft of intellectual property from research universities and pharmaceuticals by outlaw nation states, (one in particular) and the potential to hold both hospitals and their patients to ransom by just about anyone - that’s what really worries me most.

I believe we are on the cusp of an AI arms race. Attackers are busy designing new attack vectors and methods to get by cyber defenses that heavily leverage AI and ML (machine learning). Advanced persistent threats (APTs) that hide unnoticed on the network for years sometimes, while gathering vital information and gradually expanding their footprint till they own the entire network, just as the attack on SingHealth in 2017 demonstrated. AI that perfectly emulates the normal acceptable behavior of users and systems on the network and as such goes undetected by even the best cyber defenses. AI that knows when someone of significance is on vacation by their spouse’s Facebook or Instagram posts and can perfectly emulate the exact way that a CEO communicates, in order to seemingly instruct Finance to make payments to an overseas supplier from their yacht on the high seas, well out of cell phone range for any chance of voice verification.

‘Offensive AI’ mutates itself as it learns about its environment to stealthily mimic humans to avoid detection. It is the new cyber offensive weapon of choice and will automate responses to defensive measures rather like playing chess with a computer – it learns as it goes. But increasingly the intent of attacks is not just to steal information but to change it in such a way that integrity checking is impossible. Did a physician really update a patient’s medical record or did ‘Offensive AI’ do it? Can a doctor or nurse trust the validity of the electronic medical information presented to them? This is the new threat and it is best executed by AI.

Did a physician really update a patient’s medical record or did ‘Offensive AI’ do it?

Why would anyone do this? Well, I can think of at least three reasons: Cyber-war, monetary extortion, and as a distraction from even more nefarious attacks against military targets or defense secrets.

AI is already being used very effectively for cyber defense. Advanced malware protection that inoculates the LAN and responds in nano-seconds to anomalous behavior patterns. Biomedical security tools that use AI to constantly manage and secure the rising number of healthcare IoT devices as they connect and disconnect from hospital networks, (just as my company, Cylera makes). AI-powered attacks will outpace human response teams and outwit current legacy-based defenses. ‘Defensive AI’ is not merely a technological advantage in fighting cyber-attacks, but a vital ally on this new battlefield and the only way to protect us all from the cyber criminals of the future.


Q. You will be conducting a cybersecurity workshop titled “The rising threat of Internet of Things - Everything from Medical Devices to Hospital Management Systems” at the upcoming HIMSS AsiaPac19 conference from October 7-10 held in Bangkok Thailand. Could you give us a primer on some of the common IoT-related cybersecurity threats in healthcare?

A. So unlike IT devices, by and large IoT devices can’t be centrally managed, patched, updated, or secured. IoT devices are simple and functional. They open and close a set of elevator doors, and move the elevator car to the desired floor. That’s all they do. They do it well and they do it millions and millions of times during their life spans.

The same is true with medical devices that administer drugs to a patient at a certain flow rate based upon the drug library, report on vital patient statistics like BP, heart rate and O2 saturation, and scan patients for broken bones, tumors, and other ailments. Most were designed at a time long before sophisticated and well-funded nation state cyber criminals, and a time when devices were by and large not connected to the Internet. Now these devices are managed remotely from hundreds of miles away by third party vendors who can do the job better, faster and cheaper than having a number of FTEs on staff locally. Thanks to digitization and inter-connectivity, devices now communicate directly with HIT applications and the EMR – something most older systems were never designed to do. And they certainly were never designed to connect securely. By network-connecting these highly insure devices we have opened Pandora’s box, and the number of network-connected HIoT devices is growing at an exponential rate.

The big question is how do we understand what we have on our networks, assess and quantify their threats and vulnerabilities, and remediate those risks in such a way that patients are not placed at potential harm from attack by medical device. How do we identify when one of these devices is behaving abnormally so we can swap it out before attempting to treat a patient based upon inaccurate data or behaviour? How can we identify when a device has been compromised and is being used to attack the hospital? These are things that physicians, nurses, and biomedical technicians are not currently trained to look for!

The global WannaCry attack, attributed to North Korea, caused a large number of hospitals especially in the UK to have to turn away ambulances and cancel procedures. It was just the tip of the extortionist’s iceberg. Forget the de-encryption of medical records for a Bitcoin fee, just wait till patients in ICU or NICU are held to ransom - maybe by the medical devices attached to them and keeping them alive. Sound far fetched? So did putting a man on the moon in the 1950s!





Q. Cybersecurity is a constantly evolving field these days with the rapid advancement of technologies as well as the increased sophistication of cyber-criminals. How do cybersecurity professionals learn to stay ahead of the curve and keep abreast of the latest developments & training?

A. Many people who remember the 'dot com' era of the late 90s will remember the term 'Internet Year' to describe the rapid pace of change affecting IT at the time. A time where a year’s worth of development would be crammed into a few months. Well in cybersecurity, things change by the week. That includes threats, vulnerabilities, threat-actors, attack-vectors, new offensive and defensive technologies, and even a few advances on the procedural front as we discover better more efficient ways of doing things.

I can’t talk for everyone in my line of work, but I spend a lot of time reading blogs, tweets and other social media posts from experts in the field, as well as a lot of articles from the cybersecurity and industry trade press like Healthcare IT News. I also read more than my share of white papers and academic journals along with the odd book or two. My reading includes developments not just in cybersecurity but also healthcare and other industries which allows me to consider the implications of new non-security technologies and how they might impact cybersecurity and risk one day.

One thing that really concerns me right now is the exponential growth in IoT – everything from network-connected home thermostats, to internet connected refrigerators, connected vehicles, to connected cities where traffic lights are optimized to allow the free passage of emergency vehicles through rush hour traffic and everything else. This is an area I spend a lot of time researching. IoT devices already outnumber the human population of the planet, and by next year there will be in excess of 20 billion network connected devices. Now consider that even a small percentage of these devices might be out to attack you and you can see the magnitude of the problem. The growth of botnets, now far overshadows unpatched Windows machines that have been turned into zombie attack systems by their real owners – the hackers and nation state cyber forces that easily took advantage of weak security and now OWN their user’s online banking information and shady personal photographs. I sometimes think you should be required to pass some sort of drivers test before being allowed to purchase a home computer!

I also consider security and industry conferences to be a great source of vital information. I probably speak at 20+ conferences every year and attend quite a few more on top of that. I always learn something from the discoveries, war stories and experiences shared by other speakers and practitioners in the space. There’s also a lot to be learned by the way healthcare is delivered and secured in different countries even though I work in quite a few. HIMSS, CHIME, AEHIS, H-ISAC, RSA, BlackHat, and KiwiCon currently top my list, as do conferences and summits put on by various publications in the space. They are all good, and if you can spare the time and afford the admission then I find that I always come away with something new as a result.


Q. A constant challenge for healthcare organizations is the management of limited resources and budgets for cybersecurity measures, and cybersecurity can often become an after thought. What advice would you give to them in their approach to cybersecurity, particularly in light of their resource constraints?

A. In one sentence? Treat Cybersecurity risk in the same way you treat Patient Safety because the two are inextricably linked in today’s connected digital healthcare environment. Many hospital CEOs, Boards of Directors and Ministers of Health haven’t realized this yet. The sooner they do the better for all of us.

Another piece of free advice for healthcare boards is that healthcare compliance does not equal to security. The industry suffers from a myopic focus upon protecting the confidentiality of patient data, when in fact operational and reputational risks to data integrity and system availability are far more important and potentially damaging. No one is going to die because of a confidentiality breach, they could however easily die as the result of an integrity or availability cyber-attack. The healthcare industry needs to adopt a risk-based approach to security, based upon assets rather than controls or a compliance checklist. Only then, will healthcare boards begin to understand their level of exposure, and feel inclined to do something about it.

In essence we have several giant gaps currently. A gap between the ease of a perpetrator attacking a victim, making lots of money from that attack, then walking away scott-free, versus making cyber-attacks difficult and very costly for the perpetrator – whether that perpetrator is an individual, a criminal group, or a nation state. Its rather akin to the school playground where a bully is beating up and intimidating other kids stealing their lunch money, but the school rules have yet to catch up to outlaw bullying or place CCTV or a teacher in the playground to grab any bullies by the ear and drag them to the Headmaster’s office for punishment and a corrective action plan!

The other gap we have is in resourcing. According to the Cisco Annual Cybersecurity Report, there is a 12x demand over supply for security professionals. We need to train tens of thousands of security analysts, architects, threat analysts and security operations staff for the world of tomorrow. We also need to allocate much greater budgets towards securing the future of our businesses, whether that business is a profit-making enterprise or a public service. This is a simple legal question of negligence in my opinion. If those ultimately responsible choose to ignore or accept a critical risk against the advice of their security and risk executives, then they should be held liable. Especially in healthcare where patient lives are at stake.

Everyone likes to talk about the next great level of interoperability in health IT but they haven’t figured out yet that to get there, you need to invest in cybersecurity to prevent your patients from being attacked by cyber criminals and their PII and PHI stolen or altered.

Cybersecurity and protecting patients should be viewed as a “business enabler” of new more efficient, more profitable, digital health services and should be an initial design consideration not a last-minute ‘strap-on’ where you are going to spend a lot more time and money for a less secure system. “Security by design” is where we need to be.

“Security by design” is where we need to be.

A true senior security executive, is one that sits at the right hand of the CEO and frequently addresses the board on security matters. He or she directs a comprehensive holistic cybersecurity program staffed with a solid team of security professionals. Together, they facilitate a hospital expanding its range of services to patients for the delivery of more profitable services. Services like telehealth and telemedicine that improve patient satisfaction scores, and the adoption of new riskier technologies like artificial intelligence and machine learning that will ultimately improve patient outcomes by catching tumors earlier and reducing the high costs of intervention for patients with latter stage cancer or similar diseases.

No one expected the Spanish Inquisition but it came all the same

Cybersecurity will also facilitate advance of personalized medicine by protecting highly confidential information like someone’s genome sequence. A patient can change their name, their address, even their health number following a breach of information. They can’t even attempt to change their genetic sequence. Human cloning may sound rather SciFi but it’s not that far off. China has reportedly already accomplished this. In the fifteenth century, no one expected the Spanish Inquisition but it came about all the same. We need to think outside of the box to prepare for the challenges to our business model in healthcare and the threats and risks that we face.

http://pubs.cyberthoughts.org/AP19.HIMSS.Show.Daily.pdf
Click for the original Show Daily PDF


This blog was first published by HIMSS Media and Cylera

Subscribe to our periodic posts via email to new posts so I don't miss them.

Original stories and articles may be republished without charge provided that attribution is provided to the source and author. Articles written for, and published first elsewhere, are subject to the republishing terms and conditions of the host site.