The n-Category Café

Tom Hales has a beautiful analysis of the NSA backdoor to NIST and is arguing that formal mathematics (like HoTT) should be used as the new standard of rigour for such protocols.

The NSA and GCHQ must be two of the largest employers of mathematicians in the world.

The often-repeated line in the US media is that the NSA is the largest employer of mathematicians in the world.

Beilinson’s letter strikes me as somewhat overwrought. Whatever you think of the spying programs revealed by recently stolen documents, claiming that they “wildly exceed anything conspiracy theorists could imagine” is an insult to the imaginations of conspiracy theorists. (-: Of course the NSA spies on people. All governments spy on people.

Now clearly, in a free society there are conversations to be had and balances to be struck, and reasonable people can disagree about where lines should be drawn. But Beilinson seems to be objecting to the existence of the NSA, which I find absurd. Today is Veterans Day and Remembrance Day, on which we honor those who serve in our armed forces. The mathematicians and cryptographers who work at the NSA and GCHQ do not (usually) risk their lives in the service of their country, but they play a role no less important in protecting and defending our free societies, for all that most of their work happens secretly. They deserve more honor and respect in the mathematical community, not less.

I’ve just deleted two comments that weren’t constructive. Since we don’t usually have this kind of discussion here, let me make explicit what kind of thing I’m going to allow:

1. Discussion of the NSA and GCHQ is fine. That’s what this is about. Both places employ large numbers of mathematicians, and mathematics is involved in the mass surveillance programs — especially in the breaking and circumvention of online encryption. This is the relevance to the mathematical community.

2. The closer the discussion sticks to those issues that concern mathematicians, the better. If it strays too far away, I may steer it back.

3. Please try to keep the temperature down. Good ways of doing this are to provide linked references and not to appeal to emotions.

4. The obvious stuff: no insults etc. (but I hope I don’t need to say that here). I’ll simply delete objectionable comments.

The Electronic Frontier Foundation will run an exhibit at the Joint Mathematics Meetings in Baltimore in January. We could use some mathematicians who would be willing to volunteer some help. Let me know if you are interested.

I suspect that the NSA will be recruiting at the meetings in Baltimore with particular fervor.

Many mathematicians seek employment at the NSA because they cannot find work at a university.

Most of them quite frankly could probably care less about the privacy of Beilinson, or that of any other American for that matter. After all, to obtain a high-level security clearance, they had to endure a series of highly invasive interviews and background checks that can reach back a decade or more into a candidate’s past as part of building up what essentially amounts to their “FBI file.” That process alone probably does more than anything else to desensitize agency employees to the privacy expectations of others.

Worse, unlike their tenured counterparts in academia, once these mathematicians secure a job at the NSA or a related entity, they often have to suffer many of the same insults hurled upon workers throughout much of the rest of industry: rigid work hours, cantankerous and/or unreasonable bosses, meaningless meetings, unrealistic project goals, politically-driven performance reviews, restricted job autonomy, no control over one’s intellectual property, and miniscule salary increases (which often do not keep up with inflation) as the employee nears his/her retirement age. Again, finding ways to cope with such issues probably trumps worrying about the privacy rights of others for many agency employees.

Perhaps Beilinson’s energy would be better directed then toward lobbying Congress for a massive increase in the NSF’s budget that would enable more mathematicians to pursue academic careers instead of becoming a spook … rather than trying to rally them with hollow rhetoric in his AMS call to action that does little to address the underlying structural problems with research funding in this country that drives so many mathematicians to seek employment at places like the NSA in the first place.

Until then, I’m afraid that many of my comrades will continue to work for whoever can pay their bills :)

“All it takes for evil to triumph is for good men to do nothing” - John Bentham.

I do not imply by the above quote that GCHQ or the NSA are inherently evil. However, the current spying apparatus is beyond Orwellian in it’s possible future reality.

Once tools are made, machines built, systems embedded and algorithyms enacted in computer code, they could be unleashed to any end. And we do not, and cannot know, who our future politicians will be.

As many of the recent revelations show us we have been lied to already, many times over, about the status, level and sophistication of government surveillance programs.

There is a price to pay for freedom and often that price is a little risk. The status quo is that fundamentals of so-called ‘free democratic societies’ have been trampled upon: The right to freedom of expression is curtailed, the right to a private life is curtailed, the freedoms of the press are curtailed, journalistic sources can no longer rely on the integrity of the journalist to protect them - as the underlying strata of communication itself has been infiltrated and subverted.

These prices are too high for what we gain.

There is also the question of what we are being protected from and how it arises. Much of what is done and aimed for is done in the name of protection from terrorism, yet we must ask ourselves ‘what creates terrorism?’ Terrorism does not arise from hatred of freedom but the opposite, in my opinion. Terrorism arises where people are disenfranchised by the system, government, economic and political policies: disenfranchised to the point where they feel the only way to regain a voice is through horrific criminal violence. This is abhorrent but the truth is that the current (and future possibilities) of the surveillance state only further disenfranchise, as stated above through curtailing freedom of expression and presssprotections.

There is no excuse to be a good man who stands by and does nothing. The extrapolation of where we are now, if we continue along the same road, is more terrorism, more disenfranchisement, less freedom for everyone and eventually a fascistic state apparatus at the behest of media-manipulated pocket-politicians in the hands of, and made by, giant corporations.

Freedom has a price and it is a price we must wisely choose to pay if our children and great grand children are to live in a world where even the concept of freedom still exists.

This price is non-compliance with furtherance of the surveillance state and working actively to roll back some of what exists to free freedom itself once again from the possibilities of absolute tyranny.

Our children and grand children and future generations will not forgive us for idly standing by and watching the status quo unveil in it’s inevitable direction towards Orwellian totalinarianism.

A massive organization with unlimited funding, cloaked in secrecy that collects information on you but whos members dont respect the freedoms or laws of those they surveil while at the same time are caught repeatedly lying to those it was created to protect should be feared regardless if that plays into their goal. The fact that any dissent inside this organizations ranks, whistle blowers etc., will be met with indefinite jail time meted out by secret courts rather than applauded for uncovering it’s crimes, should make it quite clear that this organization is incapable of changing its path from the inside. We maintain our freedoms buy standing up to tyranny. There is no other path. Unless you think being told 2 + 2 = 5 is the other path.

I do not expect everyone to agree on the proper liberty-security balance. One’s political positions on these matters often spring from basic moral assumptions that are difficult to change. But regardless of all that, there is one aspect of the Snowden revelations that should be deeply disturbing, especially to mathematicians.

No, I am not referring to the collect everything mentality or PRISM or spying on heads of foreign governments. I am referring instead to some of the specific means the NSA and the GCHQ use to get around encryption.

We now know that the NSA does not merely attempt to break existing encyption, it actively subverts them. It does so by building backdoors into all encryption programs it can get it’s hands on. It does so by covertly and overtly influencing and weakening the international standards that govern all the protocols that run the internet (such as making sure the widely used implementations have backdoors and that only weak curves are used for elliptic curve cryptography). And perhaps worst of all, it is likely that the NSA has influenced Intel and others to weaken the hardware random number generators that sit inside your or my computer.

Not only do the above activities make the net more insecure for everyone (the NSA’s subversion of the code that allows it to crack SSL, RC4, VPN etc. in real time also allows anyone else to do exactly the same), it also destroys something intangible, namely trust. I am referring here not just to trust in important institutions but also to our trust in the ability of mathematics to make a difference in the real world. When all this perfect and beautiful mathematics is sabotaged in ways to make it useless for most of their intended applications, it is fraudulent and criminal in ways that goes far beyond what should be acceptable, even by government spy agencies. No one knows any more what programs or implementations or hardware are backdoored. The only realistic solution is to throw away a lot of work and start over.

Yes, governments spy. But some methods used to spy are more unjustified than others.

It evidently takes us mathematicians a bit longer to think these things over than the average citizen, but hopefully that might mean we can do a good job of contributing at the point where general interest begins to subside…

I actually have been putting in a lot of thought, and I wrote up an opinion piece for the Notices that is scheduled for the January issue. It is not nearly as exciting as the letter this month. However, I still have time to revise, and I’d like to take the opportunity of the current discussion to ask for suggestions.

I think abhisaha’s point above is a good one. The NSA revelations have deeply shaken the crypto community because it is no longer possible to know what one can trust. Even if there exists a good algorithm, there is a hopeless in the knowledge that it can never be implemented as wished. All the goodwill the NSA had from mathematicians and cryptologists is gone and the feeling left is one of betrayal. The good thing is that it will probably lead to a thorough check and overhaul of a lot of existing software in fact the process has already started.

I find the “price of freedom” rhetoric completely missing the point: It is as if some financial consultant would confuse normal taking of benefits (even if maladjusted to the costumer’s gain) with simply stripping his customer’s bank account and call it “price of profit”. The sort of generic voyeurism which is the topic is just incompatible to democracy. That is well known, e.g. Susan N. Herman in the chapters on the librarian case in “Taking Liberties The War on Terror and the Erosion of American Democracy”.

The selling argument for a robust intelligence agency is that politicians make bad decisions when they don’t understand the problem, and a neutral well informed source of information can prevent at least the ignorance part of the equation. Since politicians ideally serve the social compact of the governed, the more well informed the general public, the better.

This is what attracted ethical mathematicians and I have known many. The incidental fact is that some insights are gleaned from spying, and that special mathematical skills are required to do that.

But vastly higher and deeper mathematical insights are required to make sense, to understand the mathematics of logic and reasoning about cause, to ensure that something like meaningful insight results. The primary mission of the intelligence apparatus is thus, ‘to produce intelligence.’

Actually. Where the NSA and its spinoffs went wrong is in losing this mission. (By whom and how is irrelevant here.)

There is a difference between understanding a situation and finding a bad guy. These were different missions once upon a time, and so enforced under Church rules. With the former, the fewer secrets, the better the mission is accomplished. The more the public knows, the more we can keep our politicians on track.

The latter is police work, the business of another group. And we do have rules-that-matter for that. We can enforce those rules. It won’t be hard.

I suggest that the best people to influence the direction of these agencies are the mathematicians inside (well, mostly in contractors) who empower the whole apparatus. More engagement rather than less is what is needed.

A good test (to me) of whether it makes more sense to influence an established organization by a) engaging with it, or by, b) actively working against it by attempting to weaken it’s power from the outside is to check two things. One, how much of the organization’s activities, interpreted broadly, should even exist. Two, how likely it is that an ethical individual can join the organization and influence it for the better, or at the very least, stick to only doing those things within that are actually good and valuable.

The NSA, to me and many other mathematicians, fails both these tests currently. It’s mission (encapsulated quite simply as “collect it all”), is deeply screwed up and simply put, should not exist in any recognizable incarnation. And if I, an ethical mathematician were to engage with it from the inside, there is little I can do. This is a secret, hierarchical organization with little room for independent action. If I were inside, I will be part of evil (that’s how I see it, and I suspect do many others). If I, working from inside, were to dissent or attempt to make changes from within, perhaps by alerting higher-ups about my concerns, I will face the fate of William Binney or Thomas Drake.

No, the NSA is too big, too powerful and too authoritarian. It does not need to be mildly nudged, it needs to be weakened. This means countering it every way we can, by building NSA-proof encryption and email, by influencing elected officials to cut it down to size, by influencing people we know to not join it. In my mind, I can justify an ethical mathematician joining or engaging closely with Wall Street or an oil company. Not the NSA, not any more.

Hopefully this isn’t straying too far from the blog topic; there was a section in Hannah Arendts (a student of Heideggers) book Totalitarianism which I thought apposite considering the remark of Beilinsons about the KGB. The book is a socio-political study of the roots of totalitarian regimes, her main examples being both the Nazi & Stalinist regimes. Given the emotive topic, I want to make clear that I am in no way comparing current political formations with either of these: The role of surveillance is just one small strand in a complex narrative that Arendt traces over four hundred years in Western Europe on the rise of modern totalitarian regimes.

“The Okhrana, the Czarist predecessor of the GPU, is reported to have invented a filing system in which every suspect was noted on a large card in the centre of which his name was surrounded by a red circle; his political friends were designated by smaller red circles and his non-political acquaintenances by green ones…obviously the limitations of this method are set only by the size of the filing cards, and theoretically a gigantic single sheet could show the relations and cross-relations of the entire population. And this is the utopian goal of the totalitarian secret police. It has given up the old traditional police dream which the lie detector is still supposed to realise, and no longer tries to find out who is who, or who thinks what”

The interesting thing here for me (apart from the description of the antiquated technology of the surveillance apparatus) is that this is exactly the meta-data which is scooped up by the NSA.

One of the essential differences, as far as I see, is that the populace is implicated, as a large network of informers are required and which must permeate the body of the population; whereas the NSA/GCHQ surveillance apparatus is automatic. To be fair to Arendt she does forsee this possibility:

“Now the police dreams that one look at a gigantic map on the office wall should suffice at any given moment to establish who is related to whom and in what degree of intimacy; and theoretically, this dream is not unrealisable although its technical execution is bound to be somewhat difficult”.

The ultimate responsibility for the NSA spying lies with the US government: the president, the Congress and the courts.

The cause why those institutions became incompatible with democracy is the deterioration of journalism, which worked in the past because communication with journalists was safe and in democracy even protected by the law. So I think that the AMS could even offer cures (instead of only cutting the links with the NSA and other such institutions) for restoring a democratic process by mathematical expertise for rebuilding such communication routes (like antisclerotic drugs reopen veins). An absence or onset of opposition and interference would tell if and to which extent Beilinson is correct in his assessment.

Please tell me we’re not conflating the NSA with the KGB or any of its predecessors or latter incarnations. The US provides what is by any measure an unprecedented degree of freedom to its citizens, whether born here or abroad. The fact that Beilenson conflates the two is depressing, but reminiscent of the political naivete of some other brilliant mathematicians.

Instead, why not just abolish the NSA? http://petitions.moveon.org/sign/abolish-the-nsa?source=c.url&r_by=9496530

I can’t speak for others, but for myself, Beilinson isn’t conflating the two, but pointing out a salient fact that mass surveillance was a characteristic of (some) totalitarian regimes, and using the KGB as shorthand for this. This is not to say that Western democracies have suddenly overnight become totalitarian.

I don’t dispute that a state requires policing and some of this will be in secret; but politically-philosophically speaking how far must the policing of the state go before it becomes a police state? Surely this is a question worth asking when it takes only a little reflection to see that technology has progressed to a point where mass surveillance is a relatively straight-forward proposition, as has been proved the case; and given the close connection between policing and surveillance.

It might be worth quoting here from Agambens ‘State of Exception’. He is a prominent Italian Philosopher in the continental tradition.

“In a technical sense, the Italian Republic is no longer parliamentary, but executive. And it is significant that though this transformation of the constitutional order (which is today underway to varying degrees in all the Western democracies) is perfectly well known to jurists and politicians, it has remained entirely unnoticed by the citizens”.

That the indiscriminate surveillance of the public has come to pass without public and parliamentary oversight or debate surely underlines to some degree his point?

It might also be worth noting that it was a journalist with training in constitutional law that broke the story.

Mozibar, I’m in agreement with the beginning of your response, but I would strengthen it. The United States is not remotely comparable to the former Soviet Union or Russia today, where journalists and dissidents are murdered with impunity at the behest of the government and homosexuals and other minorities, particularly those hailing from former client states, are persecuted. The only recourse in defense of Beilenson’s stance is a kind of ‘slippery slope’ argument in which NSA betokens a kind of inexorable slide toward fascism, which I do not find credible. The U.S. is far from perfect, but it is also far from your example, Italy, which is beset with far more corruption at the government level than the U.S. We are not moving toward fascism; in fact, a sober look at history reveals we are headed progressively toward greater and greater freedom here, irrespective of our gender, sexuality or place of birth.

The NSA is akin to nightmares that can be totally destructive for you (and those nearby). If you don’t see them as a brain disease that serves no purpose and does not answer to anything in the outer world, the true reality, whatever it is, stops to exist for you.

Madness plays a major role in human history, no less than in personal stories. Sometimes it becomes the main force. I would think this is what happens now with the US. It is sad fact that madness ultimately leads to unlimited violence (take Hitler’s Germany, the Lenin/Stalin Russia, or the USA of XIX century).

There is Breugel’s painting “Mad Meg”: a woman, sword in hand and stare fixed, in her rush to plunder brings the Hell into existence. The NSA is a facet of the Mad Meg’s reality. Others are wars waged for the sole reason to keep going the enormous army machine, the financial system dominated by structures that enrich themselves enormously on machinations with non-existent money, etc., etc., etc. A source of it might be “The Snow Queen” splinter in our eye - the belief that the Earth is our species’ property, open as such for rape, plunder, and destruction.

It is strange that a comparison of NSA with KGB can be seen as any surprising. I’ve lived in the SU for the half of my life, and I do not see any substantial difference between the Brezhnev regime and the Bush-Obama one.

Unfortunately, the simple things are most difficult to convey, especially to truly smart people. Awhile ago I told to my old friend (who is a most brilliant mathematician) how sad it is that his son, who just graduated from MIT, went to Wall Street; he did not understand. Would it be possible to explain to Heidegger of the beginning of the 30s that joining the Nazi Party is a bad idea?

Roy,

A sober look at history whilst contemplating the contemporary condition is of course the correct position to adopt; but one may ask why are these comparisons to the former Soviet Union being made?

If someone was to ask me last year for a state that indulged in mass surveillance of its citizens I would have picked on the former Soviet Union as I suspect many if not most would do so. I’d suggest that it was the rhetoric of the cold war that established a symbolic link between mass surveillance and oppression in the Soviet Union which one supposes why it is disquieting that the NSA/GCHQ is also indulging in this.

Your own description as to why the USA is not the Soviet Union does this - you point out that it doesn’t disappear dissidents and journalists as the Soviet Union has been known to do so. If in (some counter-factual world) the USA did do so, would then mitigating factors be found so that moral distance remains?

Jeremy Scahill, An American Journalist has this to say:

“Under President Obama the US justice department has authorised the seizure of the phone records of journalists; they are tracking the meta-data of journalists; they’re prosecuting whistleblowers in record numbers under the Espionage Act; there really is a war against journalism. For everyone who does this kind of work, where you’re taking on powerful institutions, the responsible posture to take is to assume they’re monitoring your communications. It’s a part of doing this work.”

Agamben does discuss fascism in its proper context as a particular political ideology of its time with its locus in Italy; he doesn’t make the mistake of equating every concentration of power in the executive as fascistic; in fact as the late American Political Scientist Clinton Rossiter points out in ‘Constitutional Dictatorship’ this form of political power has constitutional backing, generally in times of crisis, in the four jurisdictions that he examines - the UK, Germany, France & the US. He says:

“It is written in frank recognition of a dangerous but inescapable truth:’No form of government can survive that excludes dictatorship when the life of the nation is at stake’.”

He also writes in the last section:

“Professor Rappard has pointed out that the concentration of power in the executive, the governmental invasion of the field of free enterprise, and the increasing encroachment of the state upon the liberties of its citizens are changes characteristic of all modern democracies”.

This forshadows Agambens point. He goes on to say:

“There are those who believe that these twentieth century departures of democratic government will aid the cause of human freedom. Nevertheless, all three are clearly repugnant to the Western democratic tradition; any particular step in their direction should be instituted as a permanent policy only by the elected representatives of the people and only after full deliberation and popular acquiescence, not in the confusion of some national crisis”.

I would speculate if the mass surveillance of its own citizens had been proposed by the executive to the legislature to win ‘popular acquiscience’ it would have been thrown out.

The above theoretical exegesis is underlined by what has happened on the ground in the US as the report in the New Yorker by John Cassidy points out:

“In retrospect, whats most notable about the order President Bush signed are the restrictions it contained. Originally, it lasted for just thirty days, and was limited to online communications in which at least one of the communicants was located outside the United States. Moreover, it was explicitly based on ‘the Presidents determination that after 11 September 2001 terrorist attacks in the United States, an extraordinary emergency existed for national defense purposes.’ Over time, though, the ‘extraordinary emergency’ was deemed a permanent state of affairs, and the scope of the authorisation broadened until, eventually, it came to include even the collection of data from communications between American citizens inside the United States.”

Sasha Beilinson’s last post helps put the NSA debate within the appropriate larger framework, one we can call “responsibility of mathematicians”. In the previous century, the physicists were the most powerful scientists. This power came so quickly to them and in such a perilous times (WWII), that many did not pay attention to the responsibility that comes with the influence. When one day, they woke up and realized that they have given their government unchecked powers to literally destroy the world many times over, it was already too late.

Mathematicians are arguably the most powerful scientists of the current century, because of their enormous influence in the financial industry and the NSA (and the military/industrial complex, in general). They had a huge role in the recent financial meltdown; however, I have not seen a serious debate among mathematicians about their role and responsibility in that disaster. It is a welcome change that a debate is starting to happen amongst mathematicians regarding the NSA debacle.

While I sympathise with Sasha’s call to severe ties with the NSA (or the Wall Street) I don’t think this is a very realistic solution. As long as NSA and Wall Street are socially acceptable institution, it is difficult to convince general mathematicians that they need to make a break with them. The difference with KGB is that the latter, as far as I understand, was not a socially acceptable institution. It was feared and loathed by the general population which included the mathematicians.

Sasha explains that NSA is no less effective in destroying personal freedoms than KGB. Due to my ignorance, I’m not in a position to judge this proposition; However, it is clear that the KGB and NSA don’t have the same image among the population of SU and USA, respectively. Until that changes, mathematicians need to find other ways to raise awareness and bring checks and balances to the system.

One point just occurred to me is that even with perfect encryption of content meta-data is still available.

Unless of course one can encrypt meta-data, but that given my fairly trivial knowledge of how the internet works, seems infeasible; but I could be very wrong.

I’ve just updated the post as follows:

I have now heard back informally from someone who was at the latest LMS Council meeting. (The Council is made up of academics and “is responsible for determining the strategy and policy of the Society”.)

Apparently, there was unanimous agreement that the LMS should at least be transparent about this, and should state publicly what connections there are between the LMS and GCHQ. (For example, GCHQ part-funds LMS instructional courses for graduate students.) I don’t know whether there was agreement on anything else, as I haven’t had an official response yet.