On Mon, 11 Mar 2019, Rich Kulawiec wrote:
This is why the service(s) should use confirmed opt-in on a per-device basis and offer sufficient granularity that alerts are only sent to the people who need/want them on the devices they need/want them on.
Other than nerds, which means people on the NANOG list :=), few people like configuring lots of individual devices. They usually don't. Its like blaming people for choosing bad passwords and not configuring devices securely. Defaults matter. That's why I keep emphasizing the role of "Intelligent Assistants" in these smart device ecosystems.
Apple Siri, Amazon Alexa, Google Assistent have positioned themselves as the entertainment and information device content management systems in the smart device world. When you connect a new smart device into you choice of intelligent assistant, all your emergency alert preferences should carry-over to the new device. If you turned-off emergency alerts in your intelligent assistant in the past, alerts would be off on new devices. If you want alerts on in one room, and off in a different room, talk to your intelligent assistant. You shouldn't need to remember to do that each time you buy a new smart TV or smart speaker.
10 years ago, I might have said AT&T or Comcast instead of Amazon and Google, because Cable and Telco ISPs were trying to be the home network managers. But CableLabs and ATIS have failed in that regard. Facebook is still a potential powerplayer, but seems to have missed the smart device/intelligent assistant boat.
This will eliminate some of the alarm fatigue as well as reducing the transmission requirements. It's just a rather straightforward exercise in database management.
The opt-in versus opt-out decision is a huge debate in the emergency management world.
Less than 15% of people actively opt into emergency alerts on any system, but they complain loudly after disasters. Its a bit like asking people to remember to turn on airbags or anti-lock brakes in their cars. Normal humans don't think about safety systems until after its too late. A plan to have a security guard unlock the fire exits in case of fire is a bad plan.
That inevitable human failure is why cable TV was forced to add emergency alerts in the 1990s, after a series of tornado outbreaks across the midwest, and cell phones were forced to add emergency alerts in the 2000s after a different set of disasters.
Generally, I think imminent danger warnings should be enabled by default, with easy opt-out available. Advisories and Informational alerts, such as Be On The Lookout (i.e. amber, blue, silver, etc) advisories should be opt-in, with do-not-disturb by default. Informational alerts should not alert by default, unless the user actively opts-in; and should just appear in my daily headlines, timelines, guide, whatever your smart information content manager uses.
However, just like advertisers and social media company privacy policies -- I wouldn't trust Facebook to honor emergency alert settings, emergency managers tend to ignore their promises and user preferences. Strong emergency management guidance and training of emergency alert originators is also needed to avoid alert fatigue. Its not strictly a technical problem, the people problems are harder to solve.
