What Fukushima Disaster Taught Me About Risk Management In Cybersecurity

On March 11, 2011, a massive earthquake ripped across the northeastern coast of Japan. Terrible as it was, the tsunami that followed led to an even more terrifying event: the crippling of the Fukushima Dai-ichi Nuclear Power Plant. The release of radiation, meltdowns and evacuations that followed prompted a massive investigation into what went wrong.

I was appointed chief technology officer (CTO) for the Fukushima Nuclear Accident Independent Investigation Commission (NAIIC), an ad hoc body reporting to the national legislature. The job gave me a unique opportunity to examine this catastrophe in detail and see its multiple causes. Looking at the long chain of errors and misjudgments behind it made me consider security and risk management. How could the risks have been generally discounted and so ineptly managed? What steps were taken after the accident to limit damage and what systems were established to prevent a similar situation?

After the disaster, my perspective had changed and I began to view the whole field of IT security through the broader lens of risk management. I realized that talking about security from a purely technical perspective misses the big picture. My primary task had been to design a new, secure communication system and collaboration platform for the NAIIC team of close to 150 personnel. My background in entrepreneurship and cybersecurity would certainly be tested.

Building a secure workspace

The NAIIC was Japan’s very first independent accident investigation authority established by the legislature. It was met with a lot of skepticism, and I had to push bureaucrats beyond their comfort zone to create a new entity and new ways of doing things. One of my biggest goals was maintaining the cybersecurity of the investigation with outsiders who were suddenly thrown into this activity, where IT literacy was all over the map and there was no time for education and security training. To speed the workflow, I had to allow people to use familiar tools and applications such as printing, e-mailing and file sharing. I also had to assume we going to be criticized by media and other interests and that people would naturally be careless and misplace USB sticks, lose laptops and forget cell phones. While investigating a disaster, I was trying to create a resilient system from scratch on a very limited budget. Ultimately, we were able to create a very secure working environment that used everything from digital certificates to encrypted laptops, while maintaining a high level of efficiency. Even more importantly, when data loss incidents occurred, people quickly reported them – this helped create an atmosphere of trust, which strengthened our security.

Disasters are severe teachers

One moral of this story is that it’s possible to do security on the cheap without sacrificing usability. But implementing IT security is not enough. It misses the critical component of risk management. Real security lies in maximizing our field of view and expanding our thinking.

Not being a nuclear safety or risk management expert, I also tried to contribute to the NAIIC by studying historically significant disasters as varied as the sinking of the Titanic, the Challenger explosion, nuclear crises at Three Mile Island and Chernobyl, and many others. In the Titanic sinking, for instance, a raft of failures apart from collision with an iceberg – from inferior construction methods to equipment shortages to management and regulatory errors – has been blamed for the loss of over 1,500 lives. But what I realized was that all these catastrophes had one factor in common: all came with tell-tale signs. Managers had tried to achieve a false level of “perfection,” and in the process losing valuable time and a thorough grasp of the big picture. In each case, the relevant engineers saw the potential for problems and warned their superiors, who in turn dismissed warnings due to normalcy bias.

Normalcy bias has been described in studies of disaster psychology as an unwillingness to recognize the urgency of a crisis or acknowledge that a crisis could happen. It can manifest in situations of survival as well as planning for the worst. In the September 11, 2001, terrorist attacks in the United States, 90% of survivors did not immediately evacuate the World Trade Center after it was struck by planes, instead choosing to save their work, shut down their computers or go to the washroom, according to a three-year academic study.

The lesson of Fukushima

In the case of Fukushima, warnings had been issued and ignored, and a catastrophe ensued. The commission concluded that the plant operator, the government and regulatory bodies “all failed to correctly develop the most basic safety requirements” for such a disaster. What Fukushima taught me, and what I continue to remind myself, is that in terms of cybersecurity, Japan does not have a monopoly on this type of culture. People will always take shortcuts. That’s why we must take to heart the a priori precepts of risk management: a) people make mistakes, b) machines eventually break down and c) accidents inevitably happen.

With that in mind, some lessons we all must learn as we grapple with how to prevent cyber attacks and manage risk in security:

We can’t spend forever trying to make things perfect. We waste a lot of time and miss the forest for the trees.
Murphy’s Law is a constant. Resilience is key to dealing with those precepts defined above.
We can’t change certain human behaviors. People aren’t perfect, and they make mistakes. Humans step in mud, trip over their shoes, forget their car keys – and click on malicious links in e-mail and websites. Security must be human-centric.
We can’t have an environment where an over-abundance of PEN testing and anti-phishing exercises slows productivity to a crawl. We overcome unhealthy paranoia by making security first and foremost about enhancing resilience, and then by making it integrated, transparent, automatic, efficient, prevention-focused and cost effective. We educate humans but we make our systems resilient enough to protect us from our humanness.
If resilience is the starting assumption, real risk management becomes a challenge in how best to respond and recover from all types of accidents, breakdowns and system failures, both foreseeable and as yet unimagined, by taking action at the earliest stage and assessing what is preventable next time. Hopefully, none of us will have to deal with a Fukushima-level incident, but with the growing dependence on ICT in our society today, these lessons become increasingly important for cybersecurity, as future disasters initiated or catalyzed by a cyber event are very likely.

Originally posted: Forbes
Date: 4/20/17

Posted by whsaito

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

William H. Saito