What our emergency-alert system could learn from my email provider

OPINION: Alerts may have to be issued by frightened people via disrupted communication networks — so the system has to be simple. But it also has to be idiot-proof
By Matt Gurney - Published on Jan 15, 2020
On Sunday, Ontarians woke up to an alert warning of an unspecified issue at the Pickering nuclear power plant. (Darren Calabrese)



The people of Ontario have been promised a full investigation into the rude — and frightening —wake-up call many of us got Sunday morning. The emergency alert warning of some kind of crisis at the Pickering nuclear power plant, near Toronto, was, as we now understand, an error. It should not have been issued. Reporting by the National Post suggests that the situation was a bit more nuanced than people initially thought. There apparently was a minor technical malfunction at the plant on Saturday night. But it was unrelated to the issuing of the public emergency alert, which was due to human error. The minor technical issue simply served to slow down the process of revoking the alert, the Post says. So we’re talking about a horrific bit of bad timing. But history is full of examples of exactly that.

Based on what little is known, including some information this journalist has been able to gather himself, the issuing of the alert was simply a flub. Multiple sources within the Ontario government have, in tones of considerable exasperation, explained that the emergency alert system is regularly tested — several times a day, in fact. But the system lacked proper safeguards, allowing a simulated alert to be broadcast to the public.

That may seem difficult to believe. But there is precedent. Almost exactly two years ago — indeed, two years less a day before the incident on Sunday — the people of the state of Hawaii received an emergency broadcast message from their state government warning of inbound ballistic missiles. If such a warning were broadcast to the people of Ontario, I suspect a good percentage of them would stare at their phone in bewilderment, trying to remember what the hell a ballistic missile even is. The people of Hawaii live in very different circumstances. The state is home to Pearl Harbor. At any given time, considerable United States military assets can be expected to be docked there, and it’s home to facilities essential for the operation and sustainment of U.S. Navy fleet units. The fact that Pearl Harbor was infamously bombed by surprise during the Second World War only compounds local awareness of its vulnerability and importance. Further, the state’s geographic location means that it would be the first part of the country to come within range of missiles fielded by the bellicose North Korean regime. The prospect of a nuclear attack on Hawaii is remote, but real.

The circumstances that resulted in the issuing of that alert have never been made entirely clear. There appear to have been issues with a single employee at the state’s emergency-management agency. Little has been publicly said about that, perhaps for privacy reasons. But what was immediately obvious was that a single person could trigger a statewide alert.

It seems very possible that Ontario’s error resulted from a similar kind of vulnerability. The public emergency-alerting system is not an Ontario-specific program. The system is nationwide and is designed to permit warnings to be broadcast only to those parts of the country where they would be of importance or use. Theoretically, the system could trigger a nationwide alert for an incoming enemy air attack as easily and effectively as it could warn a small, rural community of the possibility of damaging winds during a thunderstorm.

This is an important capacity for Canada to have. We are not a country lacking in geography. The ability to rapidly notify people of a variety of dangers is important. Also, such a system must be biased in favour of getting messages out, not of making them difficult to issue. It wouldn’t do the public any good if, during an actual emergency, the system were so complicated and cumbersome that it could not broadcast urgent information in real time. Emergency alerts cannot be issued at the speed of bureaucracy. Any such system must also accept as a given that, in an actual emergency, the alerts may have to be issued by shocked, frightened people via disrupted communication networks. It’s got to be simple.

But we can’t accept false alerts going out to 40 per cent of Canada’s population. An hour or two before sitting down to write this article, I clicked the wrong button on my email client, nearly deleting a 10-word note I was attempting to send to a colleague. A little prompt popped up on my screen asking me to confirm that I did indeed intend to delete the single-sentence message. In fact, I did not. I clicked cancel and sent the email. That tiny little extra step saved me the briefest passing moment of self-inflicted irritation and 15 seconds of additional work. Perhaps something like this ought to be built into our emergency-alerting system? If something like it already is, perhaps it needs improvement?

As Ontarians wait for a full explanation of the incident, we should be reasonable. As always, there may be mitigating factors here we do not yet know about. But it is important to make sure that we learn the appropriate lessons from this, lessons that we perhaps could have learned after the Hawaiian debacle two years ago. An emergency-alerting system is going to work only so long as the public has faith in it. Incidents such as Sunday’s false alarm are corrosive to public confidence. Getting the system in place and operational took many years longer than expected, and it has been beset by controversies and technical glitches ever since. It’s clear it still needs work. Let’s make sure that the next update to the system that might one day be called upon to save all our lives includes the same kind of safeguards that prevent sloppy mouse clicks from trashing our work emails.

Related tags:
Thinking of your experience with tvo.org, how likely are you to recommend tvo.org to a friend or colleague?
Not at all Likely
Extremely Likely

Most recent in Opinion