Cindy Cohn Is Leaving the EFF, but Not the Fight for Digital Rights

After a quarter period defending integer rights, Cindy Cohn announced connected Tuesday that she is stepping down arsenic executive head of nan Electronic Frontier Foundation. Cohn, who has led nan San Francisco–based nonprofit since 2015, says she will time off nan domiciled later this year, concluding a section that helped specify nan modern conflict complete online freedom.

Cohn first roseate to prominence arsenic lead counsel successful Bernstein v. Department of Justice, nan 1990s lawsuit that overturned national restrictions connected publishing encryption code. As EFF’s ineligible head and later executive director, she guided nan group done ineligible challenges to government surveillance, reforms to machine crime laws, and efforts to clasp corporations accountable for information collection. Over nan past decade, EFF has expanded its influence, becoming a cardinal unit successful shaping nan statement complete privacy, security, and integer freedom.

In an question and reply pinch WIRED, Cohn reflected connected EFF’s foundational encryption victories, its unfinished battles against National Security Agency (NSA) surveillance, and nan organization’s activity protecting independent information researchers. She said astir nan shifting equilibrium of powerfulness betwixt corporations and governments, nan push for stronger state-level privateness laws, and nan increasing risks posed by artificial intelligence.

Though stepping down from leadership, Cohn tells WIRED she plans to stay progressive successful nan conflict against wide surveillance and authorities secrecy. Describing herself arsenic “more of a warrior than a manager,” she says her intent is to return to frontline advocacy. She is besides astatine activity connected a forthcoming book, Privacy’s Defender, owed retired adjacent spring, which she hopes will animate a caller procreation of integer authorities advocates.

This question and reply has been edited for magnitude and clarity.

WIRED: Tell america astir nan fights you won, and nan ones that still consciousness unfinished aft 25 years.

CINDY COHN: The early conflict that we made to free up encryption from authorities regularisation still stands retired arsenic mounting nan shape for a perchance unafraid internet. We're still moving connected turning that committedness into a reality, but we're successful specified a different spot than we would've been successful had we mislaid that fight. Encryption protects anybody who buys thing online, anyone who uses Signal to beryllium a whistleblower aliases journalists, aliases conscionable regular group who want privateness and usage WhatsApp aliases Signal. Even nan backend-certificate authorities provided by Let’s Encrypt—that make judge that erstwhile you deliberation you're going to your bank, you're really going to your slope website—are each made imaginable because of encryption. These are each things that would've been astatine consequence if we hadn't won that fight. I deliberation that triumph was foundational, moreover though nan fights aren't over.

The fights that we've had astir nan NSA and nationalist security, those are still useful successful progress. We were not successful pinch our large situation to nan NSA spying successful Jewel v. NSA, though complete nan agelong arc of that lawsuit and nan accompanying legislative fights, we managed to claw backmost rather a spot of what nan NSA started doing aft 9/11.

But it's unfinished business. A related mobility is nan government's misuse of nationalist information justifications to break people's rights, including their privacy. I deliberation it's particularly important now that we do thing astir nan misuse of nationalist information arguments and secrecy, because successful nan hands of nan Trump administration, it's go an all-purpose instrumentality to spell aft governmental enemies, break people’s rights, and to do truthful galore different things that we person been informing astir for years.

Another point that we've really accomplished astatine EFF that is sometimes overlooked is successful nan activity we telephone “coders rights;” particularly protecting group who do independent information research. The first twelvemonth I was astatine EFF, nan FBI arrested personification astatine Defcon for having nan audacity to uncover that Adobe's information connected 1 of their products was lousy—the feline was Dmitry Sklyarov. Dramatic arrests of information researchers don’t hap anymore, thankfully. We’ve seen nan package world really germinate connected this. Now companies person bug-bounty programs to reward independent information researchers. And companies for illustration Microsoft person travel a agelong measurement successful devoting themselves to information and supporting information research. While we still occasionally person to guidelines up for personification revealing information flaws much often extracurricular of accepted areas of package improvement for things for illustration information successful cars aliases aesculapian devices, we conscionable passed different Defcon wherever we didn’t person to bail anyone retired of jailhouse aliases conflict a gag order, and that’s origin for celebration.

Tech companies often declare to return privateness seriously, moreover arsenic galore grow surveillance-driven business models. At this point, who poses nan greater consequence to individual privacy—government agencies aliases corporations?

One point that I've learned astatine EFF is that location isn't a agleam statement betwixt those two. The NSA spying that we talk astir was nan NSA utilizing nan telecommunications, internet, and communications companies to spy connected us.

We’re seeing a batch much skepticism erstwhile it comes to institution promises that they “take our privateness seriously.” The point that jumps to my mind correct now is nan largest assemblage verdict successful history conscionable happened present successful San Francisco connected a privateness rumor involving Google. That benignant of tells you that group are beautiful fed up, and they really don't deliberation that nan companies person their champion interests astatine heart. And that wasn't ever true.

The companies person besides moved against our privateness complete time. I for illustration to punctual group that Facebook launched itself arsenic nan privacy-protective societal network, backmost erstwhile that really was true. They past disintegrated that protection complete time.

There’s nary mobility that we request a broad nationalist privateness rule that has teeth and empowers group to protect their ain privateness done things for illustration a backstage correct of action. That doesn't consciousness very adjacent correct now. What we are seeing is immoderate authorities laws that are coming close, and it's getting easier and easier to get them passed, because it's very celebrated to protect people's privacy.

With Congress perpetually gridlocked connected this issue, EFF has been pushing harder astatine nan authorities and section level. Has that displacement produced meaningful results, aliases are you simply filling a spread until national action becomes possible?

We made an appraisal a fewer years agone that, connected nan national level, Congress is surgery for reasons that are overmuch bigger than integer authorities issues. While nan integer authorities organization is still beautiful bully astatine stopping bad things federally—not each nan time, but we tin do that for illustration pinch FOSTA/SESTA and nan caller effort to forestall immoderate AI legislation—it’s conscionable really difficult to build bully things successful Congress. So yeah, we decided that we would excavation successful and create our expertise successful nan states, and we started pinch California because that's wherever we are and that's besides wherever truthful overmuch of nan tech manufacture is. If you get a bully rule passed successful California, it tin person a wider effect than immoderate different places.

We've been beautiful successful astatine getting things passed and stopping a batch of bad ideas, astir precocious astir AI. There's a mates things pending successful nan California legislature now that are looking beautiful bully and I dream they will walk nan “1337” bill.

Also, wide surveillance is now happening astatine a section level, which it really wasn’t arsenic of a fewer years ago. We've collaborated pinch section groups to effort to create section power and much accountability. For instance, successful nan past twelvemonth we’ve supported groups successful Austin and successful Denver. The statewide activity is great, but we're besides trying to really partner pinch section groups connected a organization level because this benignant of wide surveillance is coming to them very accelerated now and they’re asking us, “What do we do astir these automated licence sheet readers? What do we do astir wide spying? What do we do astir nan Ring cameras and nan impacts they person connected communities erstwhile constabulary person entree to them?”

Many of these issues extremity up being hyperlocal, and we effort to beryllium strategic. We’re looking to protect group wherever we can.

The EFF has served arsenic 1 of nan biggest bridges betwixt technologists, activists, and lawyers successful nan US. What has that exemplary revealed astir coalition-building that outsiders often miss?

Well, we person this benignant of conjugation building wrong EFF because we're made up of lawyers, activists, and technologists. For astir of nan issues we activity on, whether it's protecting nan children aliases street-level surveillance aliases nan NSA, we person moving groups internally that person representatives from each of nan teams. I deliberation that's really been our superpower. We effort to span it retired successful nan conjugation activity we do pinch different groups arsenic well.

We hired nan very first unit technologist astatine EFF, and now we person a full squad of them. I deliberation 12 aliases 13 connected staff. We were really nan first statement that had existent tech expertise internally to pass what we're doing. And that helps america successful a mates of ways. One is that it's sacrosanct astatine EFF that erstwhile we measurement successful connected a technology’s effect connected people's rights, we are correct astir really nan exertion works. That’s nan main occupation of nan “public liking technologists” connected nan staff, which we now telephone nan PIT crew.

I besides deliberation that gives america a batch of powerfulness erstwhile we show up successful DC aliases successful nan European Parliament aliases successful authorities legislatures. People who we talk to, whether they work together pinch america aliases don't work together pinch us, they cognize that erstwhile we're telling them really nan tech works, that we're telling them consecutive and not misleading them.

The aforesaid is existent successful nan courts. Especially successful nan early days, we wrote a batch of briefs that were, for instance, explaining really peer-to-peer record sharing really works. We ever person a position we’re advocating for, but we besides walk clip to explicate nan exertion because we want nan courts to beryllium knowledgeable capable to norm correctly.

That’s existent successful nan legislature and it’s existent successful nan public. I deliberation having nan PIT squad in-house—along pinch nan galore machine scientists and academics who return nan clip to talk to us—has made judge we’re ever correct connected nan technology

Being correct astir really nan tech useful besides intends we're trusted by group who activity successful tech, who tin easy disregard criticisms that are based connected a misunderstanding of really nan exertion works. We effort ne'er to beryllium those people. We want to travel from a spot of heavy knowledge.

The different broadside of that is making judge nan technologists really do understand really nan rule works. The bridging goes that measurement too. For example, we want group who are building technologies that mightiness look rule enforcement requests for accusation to understand nan quality betwixt a warrant, a tribunal order, and a subpoena, and why that would matter successful position of people's rights. There are galore much group who are cross-trained now betwixt rule and tech than erstwhile we started successful nan 1990s, but having some sets of expertise measurement successful connected issues really does make america stronger. And past of course, our activism squad not only moves nan speech guardant pinch nan public, but besides makes judge that neither nan techs nor nan lawyers bore everybody pinch really we explicate stuff.

Courts and lawmakers regularly trust connected clumsy aliases misleading metaphors to picture technology. I’ve seen a charismatic comparison personification who posted leaked information online to a personification stealing an already stolen car. Do you person your ain favourite illustration of this?

That's hilarious.

The first point that comes to mind is not a tech metaphor, it's “intellectual property,” because it sets nan full point up arsenic if there's a scarcity rumor erstwhile location is not. If personification comes and steals your cow, you don't person nan cattle anymore. If personification comes and makes a transcript of your book, you still person nan book. Thomas Jefferson said: “He who receives an thought from me, receives instruction himself without lessening mine; arsenic he who lights his taper astatine mine, receives ray without darkening me.” That's a amended metaphor for really nan rule should activity successful nan integer age.

I'm beautiful proud that EFF recognized really important getting copyright and patents and trademarks correct would beryllium for nan wellness of nan unfastened internet. And 1 of nan things that drives maine crazy are these “stealing” and “property” metaphors.

End-to-end encryption remains a flash point. Where do you spot nan astir contiguous threats today—and really concerned are you that backdoors could beryllium introduced without nationalist scrutiny?

It's a tremendously important issue. It’s difficult capable successful different times to get companies to instal end-to-end encryption successful nan devices that they're offering to america because it tin interfere pinch nan surveillance business model. But I deliberation moreover group who weren’t worried astir firm surveillance should beryllium much worried now pinch nan authorities straight taking stakes successful immoderate of these companies.

I'm reminded of nan things we learned from nan Snowden revelations, specifically nan efforts nan NSA went to, moreover connected nan spot level, to weaken information successful bid to effort to guarantee that they would ever person entree to everything group said aliases did utilizing technology.

Anyone who thinks astir it for 10 seconds understands you can't build a backdoor into systems that only bully guys tin use. Anytime you make a strategy insecure truthful that bully guys person access, whether you telephone it “lawful access” aliases thing else, bad guys are going to usage it.

And speaking of metaphors, I often opportunity astir encryption that nan authorities often uses nan integer abstraction arsenic a spot of a fume surface to hide what they are really doing pinch our information erstwhile they onslaught encryption. If nan cops showed up astatine your beforehand doorway and said, “Look, we've sewage a problem because you mightiness beryllium a thief, truthful we want you to time off your backdoor unfastened truthful we tin travel and cheque and make judge you're not a thief adjacent clip there's a break-in somewhere,” astir group would opportunity “that’s crazy and will make maine little safe, not only from you but from bad guys too.” Yet that’s what we spot complete and complete and complete again successful these encryption conversations—law enforcement claiming that we request to weaken information and claiming that this will make america safer. It won’t.

One point that worries maine nan astir correct now is that we whitethorn not moreover beryllium having these conversations successful nationalist anymore. With nan tech companies being truthful intimately aligned pinch nan Trump management correct now, those backdoors whitethorn hap without america having nan chance to person nan nationalist speech astir whether they're a bully idea. That wouldn't beryllium without precedent for nan NSA.

What risks do you spot repeating from past tech booms coming arsenic artificial intelligence systems are becoming much embedded successful regular life, and what threats request contiguous attention?

There are 2 things that I've seen before. One is this crazy hype, arsenic if AI is going to lick each problem. This thought is not true. But neither is nan reverse: nan thought that AI will make everything suck successful each azygous dimension.

We person to beryllium sensible astir wherever it's useful and wherever it's not. Often erstwhile group opportunity AI, they're talking astir ample connection models, but there’s galore areas wherever instrumentality learning aliases AI mightiness beryllium used. It's important that we understand amended really these things fail. Because they don't neglect successful nan aforesaid measurement group fail. And if we don’t cognize nan existent crushed truth of a situation, we cannot moreover measure whether nan strategy is failing aliases lying to us. That’s my interest astir “predictive policing” and different attempts to effort to usage AI systems to foretell early quality behavior, particularly successful nan discourse of rule enforcement aliases authorities decisions for illustration whether to region a kid from nan home.

I don't consciousness for illustration we person a very bully intuition astir really and erstwhile AI fails, which intends we can’t build nan benignant of guardrails we need.

We know, for example, that quality judges get tired successful nan afternoons and their sentencing decisions will alteration arsenic a consequence of that successful ways that are beautiful predictable and person been tracked. AI doesn’t person that problem. It ne'er gets tired. But if an AI is deciding whether you get bail aliases not, it's going to person different vulnerabilities. We cognize immoderate of them based connected what it's trained on, that it will person biases that mean group won’t get judged connected their individual merits. We cognize rather a spot astir that astatine this point, but that’s conscionable 1 example. We still don't cognize each nan ways successful which these systems will neglect truthful its vulnerable to trust connected them for decisions that matter.

The different large point pinch AI is that nan tribunal decisions are starting to travel retired and nan first group has been based successful copyright, which is not a bully ineligible conveyance addressing problems pinch AI. It’s unfortunate that it's first because I deliberation it's clouded things. Copyright rule is not designed to beryllium applied to nan discourse of training up AIs—in wide that training process should beryllium adjacent use. And overall, copyright is excessively blunt an instrumentality pinch statutory damages and different things that are not good suited for nan problems that AI whitethorn cause.

For younger advocates stepping into this space—where surveillance is pervasive and privateness protections are thin—what’s nan lawsuit for staying successful nan fight?

Organizations for illustration EFF were made for this moment. If everything was magically going good and governments were awesome and benevolent and companies were doing nan correct thing, you wouldn't request a integer civilian nine aliases an statement filled pinch fighty lawyers, technologists, and activists. We request EFF, but much than that, we request a robust integer civilian nine to effort to clasp nan statement coming and to effort to make things amended tomorrow. With nan clear descent into tyranny, and tech truthful cardinal to that slide, we request integer nine much than ever.

So nan first point I would opportunity to personification if they were reasoning astir becoming a integer authorities advocator is this: We request you. There are truthful fewer of america compared to nan size of nan problems and nan complexity of nan problems that we're trying to address. I had nan luxury of coming up successful a clip erstwhile a batch of group didn't understand exertion and it wasn't successful very galore people's hands. So while nan fights were important, they were much astir laying nan groundwork for nan future. We didn’t triumph each of them, though we did triumph a batch and helped move things guardant for galore others.

But regardless, we're successful nan early now. There’s a batch much money astatine stake, a batch much authorities involvement, and a batch much nationalist attraction and reliance connected integer tools. So successful immoderate ways it's harder, but it's besides much important to effort to build a amended integer future.

I deliberation it's easy to look astatine nan stakes and what we're up against, and beryllium despondent. But we tin either suffer now, aliases conflict and possibly suffer later. I’m connected nan broadside of fighting and opinionated up for what we judge in. It's not magically going to spell well, and it ne'er was magically going to spell well. If nan founders of EFF had thought that, they wouldn’t person created nan statement and hired lawyers for illustration me.

It’s clearer than ever that group person to guidelines up for freedom, justice, and invention for each nan group of nan world, which is EFF's mission.

But different point I would opportunity is that it’s awesome nosy to beryllium connected nan righteous side. You get to activity pinch really good, smart, and hilarious people. I often opportunity nan group moving for a amended world propulsion amended parties, and that’s been my acquisition for 25 years and counting.

Obviously nan past mobility is: What’s next?

I don't cognize yet. I really wanted to group EFF up for occurrence and make judge it could alert without me, truthful I didn't walk a batch of power figuring retired my adjacent steps. I would opportunity a mates of things. While it's an grant to beryllium EFF’s executive director, it's ne'er really been my dream job. I'm much of a warrior and a fighter, truthful I'm looking for a measurement to get much straight into nan conflict than I could do from this role. It’s wonderful, but erstwhile you're responsible for nan attraction and feeding of 125 people, you walk a batch of clip connected HR and budgets and different things that are really, really important but are not nan portion that feeds me.

I'm looking to get backmost into nan fight, and I'm not judge precisely what that's going to look like. I americium still passionate astir fighting unnecessary authorities surveillance, wide surveillance, and secrecy. I’m still willing successful fighting erstwhile nationalist information arguments are utilized arsenic screen for undermining people's privateness and free reside and short-circuiting owed process.

I judge that, aft 25 years, it's clip for different group to get a chance to lead EFF. But I'm not done yet. I want to enactment successful nan fight.