One of the strange things about our current political order is that we both live in maybe the most surveilled society in human history, and yet the terrorist attacks this surveillance is meant to stop keep on happening anyway.

We know that under the NSA’s mass surveillance, the US government can look at almost everything you and I do on the Internet. We know the FBI has rampantly and illegally tapped into this database as part of its vast domestic spying operation often targeting black activists, partnering as well with private data brokers to amass a vast trove of geolocation and social media data on the US public. We know the CIA has its own legally dubious mass surveillance program that it’s operating at home. And we’ve just found out ICE has now become a de facto domestic spying agency through its access to the many, many public and business records we rack up in our daily lives. This is all really just the tip of the iceberg.

Yet once again, we have another horrific attack, this one in Buffalo where a white supremacist shot to death ten people just days after posting his racist manifesto online on Google Docs.

The devil’s bargain we were forced into demanded we trade away our privacy for the sake of security. Yet the massive database of intimate details about our lives that government agents can track and comb through seems yet again to have failed to guarantee the latter — even though this attacker had recently taunted and threatened law enforcement online and made threats to his school, prompting a visit from state police.

It’s a serious question about what purpose exactly mass surveillance programs serve. Take the NSA’s unfathomably vast mass surveillance system, for example. When the NSA’s spying powers were under threat following the Edward Snowden leaks, its former chief Keith Alexander famously claimed its surveillance had foiled fifty-four terrorist attacks, a claim soon uncritically repeated by a host of congresspeople and media outlets.

Yet when pressed, the only example the government would give of the program’s controversial phone metadata collection program actually being central to foiling a terrorist plot was that of a Somali cab driver in San Diego sending $8,500 to terrorist group al-Shabaab. Alexander soon admitted under oath that not all of those fifty-four plots were actually plots, they weren’t all thwarted, and only thirteen were actually connected to the United States.

A later analysis of 225 terrorist cases found that the NSA’s bulk collection program and the Section 215 powers under the Patriot Act played a minuscule role, that the agency’s metadata collection “had no discernible impact on preventing acts of terrorism,” and that officials “exaggerated the role of the NSA” in three key cases used to defend its mass surveillance. A Department of Homeland Security document later declassified showed the agency had played no role in foiling ISIS plots.

A 2010 study from the American Security Project similarly found that, looking at thirty-two thwarted plots since 2001, post-September 11 counterterrorism powers had been “instrumental in disrupting terror plots in only a relatively small number of cases.” Most of the attacks, it determined, had been foiled thanks to luck and public tip-offs as well as traditional, pre-9/11 investigative techniques, including informants and undercover officers. Given the long history of these types of agents manufacturing their own terrorist plots by manipulating incompetent and mentally ill Muslim men, there’s a good chance that some number of these “thwarted plots” were in fact glorified entrapment.

A 2015 working paper from the Pentagon-funded RAND Corporation cited, among other research, a 2011 study that examined 176 anti-American terrorist plots, similarly concluding “that plots are most often foiled through conventional law enforcement activities” and that “the majority of plots are thwarted using information provided by the general public or state/local law enforcement.” While general “intelligence efforts” produced clues in only 14 percent of plots, informants doing in-person investigative work were the largest source, with associations with known suspects, public tips, and discoveries during seemingly unconnected investigations following behind.

It’s hardly surprising mass surveillance isn’t helpful for stopping attacks. Even the NSA’s own experts privately complained that the sheer mass of data they were collecting led to a paralyzing “information overload.” A recent report revealed that ICE collects so much data on American adults, they contract with half a dozen companies just to make sense of it all.

What it is useful for is gathering private, intimate information about the lives, thoughts, and personal habits of high-profile dissidents and political figures. You may not be able to detect the next attack while swimming in this ocean of data, but if there’s someone whose communications, movements, or Internet browsing habits you want to take a look at — perhaps so you can find something damaging to discredit or blackmail them with, as US officials once tried to do with Martin Luther King, Jr and Daniel Ellsberg, among others — you can just search their name and pull up a handy dossier.

It’s also worth asking how it is that yet another attacker who was already known to law enforcement has carried out another attack. As the team at Breaking Points compiled, this has been a depressingly common feature of various major attacks, from the Parkland shooting and the attack on Orlando’s Pulse nightclub, to racist murderer Dylann Roof and the Boston Marathon bombing.

It’s not even a US-specific issue. All over the world, from France and the UK to Australia and Canada, terrorists who have gone on to commit heinous crimes were on the radar of law enforcement beforehand, and in countries with vast, Orwellian systems of surveillance no less. In New Zealand, the Christchurch shooter was allowed by police to get a gun license after a remarkably lax vetting, and his association with the Australian far-right and alarming social media posts escaped law enforcement attention — something an inquiry blamed on security agencies’ obsession with Islamic terrorists over other extremists.

Then compare some of the treatment left-leaning activists have gotten at the hands of law enforcement. The FBI has a habit of trying to preempt protests at political party conventions, visiting activists’ homes to warn them against it. The Bureau has relentlessly spied on Black Lives Matter protesters since the movement began, in one case partnering with a company to read social media posts in real time and find out plans and locations in advance, at times visiting activists mere hours after they posted about going to a rally. Anarchist Daniel Baker was raided and thrown in jail for little more than a series of Facebook posts and a flier. An alarming number of black and indigenous activists against police brutality have suffered the same.

Clearly, law enforcement is keeping a close eye on someone. They just happen to have a very different demographic profile to many of the men who keep carrying out attacks like in Buffalo.

Law enforcement has a legitimate role to play in preventing mass shootings and terrorist attacks. But as this most recent horror show reminds us, the increasingly invasive, privacy-trampling tools we’ve been told are the way to do it don’t help. And those using them seem to have their eyes on a very different set of targets anyway.