The BART incident prompts a lot of legal questions because we are facing questions for the first time. A government agency silenced a communications network to disrupt a political protest critical of the government. It is an area of high free speech importance, but not an area of settled law or practice. And, as many have emphasized, however the US agencies decide this question will set the bar (or the “ceiling”) for other nations.
This post is about “kill switches” and what the legal limits should be to using them. A kill switch is a term for a government “switch” to shut off a communications network.
People seem to mean something technical by “kill switch,” but I think the technical means are beside the point. BART has emphasized it did not jam signals; this emphasis results from BART reading the Communications Act to ban only the technical means of jamming signals (at least some disagree with that reading). Rather than jamming, at first, BART said that it had simply requested carriers to shut off the network. And a later update clarified that BART did not need the carriers; it could flip a switch on its own.
“How” is less important than “what.” Debating (1) jamming, (2) making requests, or (3) flipping a switch reminds me of the Robert Frost poem about whether the world would end in fire or ice. All three gets the job done. All three are “kill-switches.”
Kill switches have prompted huge debate. When Iran or Egypt throw the kill switch to stifle election or employment protests, the West is justifiably outraged. But in the US, in debates over cyber security legislation, legislators have debated authorizing the president to shut off communications networks during times of emergency. A kill switch has been controversial, largely because of civil liberties fears and unchecked discretion for the President (without congressional or judicial oversight).
I think the BART incident, while a local agency, provides a good concrete example to help us think about kill switches even at the federal US level.
On the one hand, the President generally has and should have the authority to respond to emergencies. Institutionally, he is the commander in chief and the chief executive. Authority should reside with him.
But our constitutional system is built on the assumption, articulated by James Madison, that men and women aren’t angels. If they were, government would be unnecessary. If flawless angels agreed to take over our government, we wouldn’t need checks and balances (or a bill of rights). But government officials aren’t angels; they’re people. Like the rest of us, officials are self-interested, have individual weaknesses, lack perfect information and foresight to select the best ideas, and often would lack the ability (or organizational structure) to implement those ideas. That’s why our government requires collaboration and compromise among several different branches of government, among states and federal government, and so on. It’s why we have structural checks and balances–we don’t just pass laws limiting government actions, we also build limits into the very structure of our government.
The key questions for designing a kill switch then, is to ensure that the president (or relevant state official) has the authority and means to keep the country safe during a real emergency, without permitting him/her the ability to abuse that authority. To do that, we need to define and limit the meaning of an emergency and implement structural limits on the scope of officials’ authority.
First, what’s an emergency? A political protest (even at a subway station) is not an emergency worthy of the kill switch. The law should be very clear that political assemblies, protests, gatherings, and the like are not emergencies. The BART situation was even more problematic: the assembly was to promote the BART itself. Lots of people would want to silence their critics and government often has the power to do it. That’s why the law should be clear that protests are not an emergency. In the context of cybersecurity, an emergency might be a worm attack on electricity grids or financial markets. But a protest shouldn’t be seen as an emergency.
Second, before or during an emergency, there must be structural safeguards. While the executive needs flexibility to respond, the executive also needs balances. There should be judicial oversight, if streamlined to move quickly. And there should be a role for elected legislators; an agency should have to seek legislative authority either before invoking a kill switch or within (say) 24 hours. As important as judicial or legislative involvement is public involvement. the public should be involved in determining questions of safety and liberty; after all government agencies are claiming to act to promote the public‘s safety. The public should be trusted with the facts–presented as transparently and openly as possible, subject to security concerns–permitting the public to determine if an emergency warrants shutting off the public’s means to speak and assemble through the use of modern technology. So we need judicial oversight, legislative checks, and transparency to ensure public involvement. (To BART’s credit, BART was open about its actions after the fact, enabling public debate on the actions.)
These questions–pertaining to a communications kill switch–are central to what kind of democracy we will have. The public should be involved in answering these questions. We should look both to substance–in narrowly defining the emergencies for which we authorize a loss of liberty to protect our safety–and to structural checks in line with James Madison’s Founding insight.
Marvin Ammori 