SPEAKER 1: She has won many, many awards in our area. Allow me not to go through the list of that. So I'm really looking forward to hear what she has to say about the intersection of technology and politics.
MARGARET MARTONOSI: Thank you.
Thanks, everyone. It's great to be here. This talk is really a chance for me to tell some stories from the year I spent in DC and the eight months I spent after that. Basically, think about it as an engineering nerds exchange program into a different world. I tend to be a fairly direct and unfiltered person. And so for me to uproot myself and go into a place that was less technical, and very filtered, and diplomatic was different. So this is partly about my experiences within that world, and partly about some thoughts on how we should all as engineers and as citizens engage more in these issues. And since there's been a guessing game while we were waiting from the talk to start, I will say that that is Geneva, Switzerland, where I spent a fair amount of time during the 20 months that I was working with the State Department.
So, quick introduction. So who am I? I'm a computer architect, which has nothing to do with buildings. It has to do with designing the hardware and the hardware software interfaces that go inside all kinds of computer systems. And if I wanted to define computer architecture very loosely in a single slide, I would say that computer architecture is the mediator between application trends and underlying technology trends. So we worry about things like how semiconductor technologies are improving over time. That's coming up from below. And then we look at opportunities in terms of what new applications might we want to exploit and accelerate. That's coming from above. And computer architects try to make that all work and fit together.
I've also done some mobile computing work that was related to some of this, basically motivated by power efficiency questions that we had. But as Jose mentioned, I've done a lot of different projects over my now 23 years at Princeton, and some of them are what I would call, not policy work, but policy adjacent. So as Jose mentioned, we built an energy efficient system for tracking wildlife in Kenya. The goal there was to understand their use-- wildlife's use-- of the land so that we could help guide land preservation policies, and see how that balanced against agricultural uses of that same land. Another project that we did following from that had to do with very low cost internet connectivity. It sounds quite different, but actually some of the same techniques that we use to get data transferred back from our wildlife tracking system were applied in Nicaragua to provide low cost internet in rural villages that didn't have any high speed internet and didn't have at the time any cellular connectivity. I also have looked at ways of using cell phone data to map human mobility on an urban scale to guide things like transportation plans or carbon footprint estimates. And I've done a lot of work over the years on power warrior computing, as well.
So all of these have the sense that if you jumped over the fence you could find policy issues associated with them, whether it's climate change issues or broadband connectivity and rural regions, and so forth. But none of the time was I spending deeply on the policy side. I was always on the tech side of things. But I was up for a sabbatical about three years ago and I wanted to do something different. I had done the sabbatical where you go to a company and learn deeply techie things, and I had done another sabbatical where I just hung out at Princeton and didn't teach. And I wanted to have my third sabbatical be something really different, fish out of water. And so with that in mind, I applied to this Jefferson Science Fellow Program, in which about a dozen fellows are placed at the State Department or at USAID, the development arm of the State Department, for a year. And the goal of the program is to bring in senior engineers and scientists into the State Department to act as subject matter experts. You get a little stipend. So you basically you get your university salary, a little stipend to live on. In my case, I was there full time for one year, August, 2015 to August, 2016.
And then I was remote from New Jersey working for them part time from August 2016 till about six months ago. And yes, I did have a diplomatic passport, and no, it does not give you immunity on traffic tickets. Those are two separate things. I had the passport, not the immunity. And when they give you the passport, it comes with a brochure that says, this is not immunity from parking, blah, blah, blah.
So when you do a Jefferson Science Fellowship, you get placed into an office and, in my case, I was placed into an office that had that acronym. So it was inside the Economics Bureau, inside an office called Communications and Information Policy, and within that, inside the office of Multilateral Affairs. So it's like Unix, they know how to have a hierarchical naming system. CIP plays a role. Anytime there's a computing or communications issue that has some international aspect to it, CIP tends to play a role. So one thing that we all know and love is cell phones, when we land in a new country, we can turn them on and they work. That's actually because some people every four years, from all the countries on Earth basically, spend a month in Geneva pulling all-nighters to negotiate frequency spectrum agreements for the next four years. So frequency spectrum is a biggie, but there's other things. With some fairly frequent regularity, some government is turning off some part of its internet or some applications that run on the internet. So blocking Twitter or turning off the internet entirely, those things happen. And offices in other parts of the world try to respond and encourage those governments to turn back on their internet. It also plays a role in a bunch of different what were called digital economy issues-- places where computing technology use is influencing the economy in big ways.
And a subtle undercurrent of a lot of the things that we did is wanting the internet to stay free and open, and a fear that a certain country was trying to turn it into something that countries control as a sovereign piece of infrastructure, as opposed to keeping it free open and interoperable. The other part that I'll explain is this word multilateral. So that's multi-country. So any place where multiple countries would come together to discuss these issues, that was the MA office's purview. So that's things like the United Nations, obviously, is where a lot of countries come together, but also G7, G20, OECD, and so forth. Different subsets of the countries on Earth. And then across the hall was the corresponding bilateral office where one-on-one discussions with particular countries would happen.
OK, this is the US government in one slide because often people need a refresher course, right? So the State Department is here. It's part of the executive branch of the US government. So we spent remarkably little time dealing with Congress. Actually, when I decided what fellowship to apply for, I had what I believe is the wisdom to look at the situation and say, I'm staying away from legislative branch. OK, so there are fellowships that will let you embed yourself in Congress. At that time three years ago, I chose not to do that. And within the executive branch, the Department of State does the foreign policy. So enough on that. Administrative foreign affairs is what it would be called and a lot of other countries, if that's more familiar. And then the third part is the judicial branch. In addition to the three biggies, there's a whole bunch of independent agencies of different types. One that is relevant here is the FCC-- Federal Communications Commission. So often when we went on delegations or negotiated tech topics, it would be State Department, plus the FCC, plus sometimes the Department of Commerce, as well.
OK, I promise there aren't too many org charts, but here's one more. This is the Department of State as of November 2016. OK, so secretary of state up high, and then these different undersecretaries doing different chunks of the work. And then a bunch of these special projects offices, even though they're depicted down low on the slide, they basically are stood up by a given secretary of state, and they report straight to that secretary of state. So the relevant thing on this is these three red parts are where computing happened at the State Department when I was there. So I was in EB, as I mentioned. That's here-- Economics Bureau. So that's one place. This is the cybersecurity office that was one of the S-level offices reporting straight to the secretary. And then this is Democracy, Human Rights and Labor, which is things like freedom of speech, making sure journalists have the connectivity to tell stories, and so forth.
That was then. That was November, 2016. That's the last org chart you can find for the State Department publicly, and there's been some news. So one thing that's clear is that all those S level positions that I showed at the bottom, those are likely to go. But actually that's not so abnormal because they're stood up by a particular secretary of state. They tend to refresh with each secretary of state. One of those was the head of SCCI So he's left, although I'm told that he's a, still officially on leave and not quite left, and b, he's coming to give a seminar at Cornell sometime soon. When?
AUDIENCE: Two weeks.
MARGARET MARTONOSI: OK, so this guy will be here soon. I will also say that in addition to the cyber offices that I pointed out on the org chart, there's another office just underneath CIP in the EB part of things called the science and technology advisor to the State Department-- STAS. The STAS has also left, and there's been no commitment from Secretary Tillerson on whether a new STAS will be appointed because there's a massive reorganization going on across the State Department, and so they aren't committing to anything right now. So that leaves us with a score that's something like this. Most of these will probably go away or get reformed. SCCI is likely to go away if it hasn't already. DRL is a big question mark. There's news articles you can find about it. STAS is a big question mark. The current STAS is gone. One good thing that might come of it is the best sense that we have is that the folks who are left from SECI will go up and merge with CIP and form a merged office. And frankly, based on the kind of overlap and weird divisions that I saw, I think that would actually be a good outcome. And there's some people advocating that it should come out of EB and be its own entity rather than being part of an economics-driven office.
So that's orientation. So I think one thing that surprises people is there's computing people at the State Department. It doesn't seem like a very technical place compared to, say, the FCC or the Institute for Standards and Technology. But yes, there were competing people at the State Department, and overall, across not just computing but all the sciences, there were hundreds of PhD level scientists at the State Department. If you think about things like public health, climate change, cybersecurity, there are places where you want technically informed people working on these foreign policy issues. And so there were plenty of scientists there.
Just to give a timeline-- so February 15 I was named a JSF, matched with an office in July, started in August. It was pretty weird to actually agree to do a program without any commitment or promise about what I would actually be doing. I mean, as late as July, I was interviewing with an office where I would have been part of the Australia and South Pacific desk, and just visiting islands in the Pacific and encouraging them to get better connectivity, which would have been cool in a different way. And then what you can see is a sequence of these meetings that I'll talk about more. August, 2016, back to New Jersey. A couple of more meetings. And by April, 2017, I decided to cut short my one year extension and get out of town mentally.
So one of the things that people-- when I say I was at the State Department and I'm a professor, they have this mental picture of me in an ivory tower somewhere, thinking deep thoughts and expounding my wisdom on people. And the best picture I have for how it really was at the State Department is that I was one of the people on these oars, pulling the boat along, and the joke was like, let's just hope that all the oars are going in the same direction. I think in general, the government is filled with skilled and dedicated civil servants who are really trying hard. And a lot of time, we were all on the oars and it was going in the right direction. And that was actually one of the most heartening observations I had of my time there. There wasn't a lot of expounding from high towers, where I was anyway. It was a lot of sort of working on things in the trenches and just keeping things rolling. So this is the mental picture of what I did. And I think it was useful to understand in a nuts and bolts way how government stuff gets accomplished, and how a lot of foreign policy negotiations are happening. It doesn't just sort of get revealed in Paris one day when it's a done deal, but it's years of very frequent meetings.
So what were my roles? So you can think about what I did in terms of the topics that I looked at, and you can think about it in terms of the roles that I had. So the topics that I looked at were a whole lot on Internet of Things, a bit on technical standards and how they influence economy and trade, a bit on FinTech-- blockchain and bitcoin were just sort of starting to move up the hype cycle as I left DC, and so I followed that for a while-- and a bit on internet governance issues. Across all of that were things like security and privacy, artificial intelligence, quantum computing came up a little bit, as well. I can tell a story about that in a moment. In terms of roles, a lot of what my role was to be an explainer, to write briefs on particular issues, and to answer questions when they came up. I did get to be, quote unquote, "on the microphone" doing negotiations. Not like high-level, tense nuclear weapons kinds of things, but negotiating different issues related to the topics up there.
So I would say that these two were Margaret being well-behaved and doing what she was supposed to do and earning goodwill, and then I would try to use that goodwill to get them to do new stuff. And frankly, getting them to do new stuff is harder than playing a role in existing stuff, not surprisingly. Some of the things that we were trying to do were to find countries that at the time were changing dramatically their relationship with the US, and then make a case for why computing and entrepreneurship could be an interesting sort of first set of events to do in these new countries. So Cuba and Burma or Myanmar were two examples of countries that were coming out of sanctions just as I got to DC, or in the first months that I was in DC, and so advocating for ways to improve engagement with them was part of what we worked on.
So what did I do all day? And this was actually a great mystery to me until I landed. Like, what actually do you do all day? I would say most days I was either writing or reading. So what kind of writing? A lot of writing talking points so that someone else who didn't know an issue as well would have a bulleted list of reference points. A lot of info memos on different topics. My classic story here is folks who are reading about quantum computing. They were reading about other countries that seemed to be ahead of us on quantum computing, and so they wanted an info memo about quantum computing that would go all the way to the top of the org chart. And I heard that they were going to write it-- they, meaning two life scientists-- and so I said, hmm, maybe I could help with that. Maybe I could take first crack at writing that draft for you.
So I wrote a draft. I put a lot of effort into this. It was four pages long. I was so proud of it because to explain quantum computing in four pages to people-- so the fundamental way that you explain quantum computing is to say how it's different from classical computing. Classical computing, the bits are all zeros or ones, and in quantum computing, it's the superposition of possible values in the state, right? Well, they didn't know about the bits being zero or one, so there was like half a page on that before you get into the quantum part. Ah, but, it turns out that at the time-- it's probably still true-- at the time if you wanted a memo to go all the way up to the top, it had to be no more than two pages long.
So my four page beautiful treatment of quantum computing, I had to subsample it down to two pages. And I will say that it lost something when it went from four pages to two. I'm a little regretful of that. Also, a lot of the meetings that we attended, and I'll talk more about them, involved written submissions. So you would write out the US position as a way of starting a conversation at the meeting. So that was one thing that you wrote, and then you also wrote what are called position papers, which are a response to something that got contributed by someone else at the meeting.
The converse of this is if everyone is writing a lot then someone better be reading. And since it isn't typically the higher-ups or it isn't typically very deeply technical people, I would read and interpret things that were written and try to highlight places that seem to be problematic to the US position in some way or another. So for example, the OECD is an organization that is mostly the European states. It was formed after World War II as part of the Marshall Plan, and it's broadened to include other developed economies in North America and around the world. So it's about 27 countries now, but it's more European than other.
And so they would write things as an OECD document but with a heavily European view of some issue, like how to manage technical standards or what should the privacy expectations be. And sometimes that wouldn't jive with the US point of view on that subject, and so there would be a game of marking up or commenting the draft with the US reaction to those points to try to get a draft that was more reflective of all the member countries, rather than whomever had initially written it. And in the end, a lot of explaining. So trying to be a reliable technical voice. As I said, I went to a lot of meetings. I think one of the top five most surprising things to me about that year, 20 months, was not just the need for engineers to be in the room at meetings but, really how often meetings were occurring.
So you could go to Geneva, Switzerland, and hang out at the International Telecommunications Union-- the ITU, which is the UN agency for computing and communications-- and I swear you could live in Geneva and attend meetings almost continuously. There's a few Swiss holidays that they take off, but basically, there's two week chunks of meetings on a variety of topics. And at one level, you'd say, oh my gosh, enough is enough already. But another level, if you weren't in the room, then the outcomes of that meeting wouldn't reflect your position on things. And so countries felt obliged to be the room for the discussions to make sure that nothing untoward happened.
The ITU has three subportions. The ITU-R is the radio frequency part of things. So they ensure worldwide spectrum gets used in ways that are efficient and harmonized across countries. ITU-D-- D is for development-- that part of the ITU tries to ensure that all countries get the help they need in developing up computing and communications infrastructure. So that's things like helping countries learn how to set up an internet exchange point within their borders, helping them roll out cellular connectivity, and so forth.
The third clump is called the T-sector. I don't know what T stands for, maybe technology, technical standards. For those of you who are of a certain field and of a certain age, you might remember learning about the CCITT network layering structure. I see someone of the right age and topical feel. OK, so CCITT was a previous name for the ITU-T, and it used to have relevance as a technical standards body. These days, it's a little light on technical. And so it's a place where countries come together to try to issue standards, but it's this funny thing where there aren't enough engineers in the room a lot of the time for them to be truly technical. And so it in many cases has been hijacked for more political purposes, such as in June of 2015, the ITU-T started this new study group. Everything the ITU-T does is in terms of study groups.
There was a new study group 20 on Internet of Things, and it was started over stated reservations from poor countries. And everything in the UN system is supposed to be by consensus which was sort of surprising to me. I thought, as a quantitative kind of person, that there'd be a lot of voting, but voting is anathema. We do not vote. We have to pretend that we have reached consensus. And so the fact that they would start a group despite four countries not wanting it is sort of surprising because it didn't look like a consensus to those countries.
So that's something that the US, UK, Canada and Germany are not happy about. And then right there, I land in DC and the first thing I'm assigned to cover is the Internet of Things. And so, basically, we spent a lot of time just trying to hem that in, and I can talk about that, and this was the total meeting count of where and when. So it's about four months total of time in ITU-land.
How does ITU-land compare to a conference? So when I go to a technical conference, I can go out into the hallway any time I want. I can go to the bathroom anytime I want. If I get bored, I can go home or I can go out to dinner with someone. That is not true in ITU-land. In ITU-land, you are there to cover an issue. If they start talking about standardizing privacy and the US person who is supposed to be in the room to keep them from talking about standardizing privacy is out having a drink, that's not good. So you basically have to stay there until the meeting gets adjourned to cover the issue.
And for each issue there would be a US spokesperson designated who's on the microphone sort of speaking out the US position. There's also a very regimented speaking style and set of interactions, which I found-- it just took me a long time to do it with a straight face. So when you want to speak, you're sitting at a desk like this. There's a microphone like that. There's ear phones. Sometimes there actually are live translations into the six languages of the Union. And when you want to speak, you press a button which lights up a light up at the front. The chair sees a queue of who wants to speak, and recognizes someone.
So they'll say the United States. Then you press another button to activate your microphone, and you say, thank you, chair. And if you don't thank the chair, that would be very shocking. Thank you, chair, and good morning colleagues-- because you have to be nice. We really appreciate the contribution from XYZ. It's very thought provoking, however, the United States believes that, blah. So you would have to wrap your concerns in appreciation and thanks upfront, then you'd speak for a while. And then you would say, thank you. Then, you'd turn off your microphone and the next person would do it. So if you're wondering why it took two weeks for all these meetings, it's in part because of that. Chris.
CHRIS: So as it comes to votes, were you representing the US or were you playing a supporting role and some other State Department diplomat was the one with the microphone?
MARGARET MARTONOSI: So in October, at my first meeting, I think I said precisely zero because I was still learning, not just the speaking convention which was odd, but also the rules of order for how the meetings went and what it meant to place a reservation on something. That's a formal thing that you do. And so October, no, but by the end in Tunisia, I was on the microphone and the whole bit. And actually one of the reasons why I extended my time with the State Department was because only government employees can be on the microphone here. The delegation includes industry people, as well, but industry people aren't allowed to speak for the US. Only a government employee can. And I will say that, in today's world, we're quite polarized right now, and patriotism, let's say, is something that maybe shows up on one side of the polarity more than on the other.
But as someone who went through this experience, I would have to say that to speak on the microphone for things that you believe in that happen to also be what your country believes in, and to use that wording, the United States believes blah, is about as patriotic a thing as I've ever felt. And I can talk about that more later. So there's something really fascinating and weird about speaking on behalf of your whole country, and writing on behalf of your whole country. So when you write these written contributions, it's not, I want to do this or we want to do this. It's the United States blah. Even when you use Microsoft Word track changes, you change-- so instead of Margaret Martonosi being the editor of the track changes, the United States is the editor. So, fascinating.
OK, one example, Internet of Things, but you could come up with similar. This is just a topic that I covered. These things came up in other forms for other topics. So what is the Internet of Things? The Internet of Things is this vision that we will apply computing communication technology to a range of physical objects in order to improve their behavior, analyze-- sense, analyze, and act on different bits of data that we get, whether it is improving the behavior of the power grid, autonomous vehicles, improving water quality, and so forth. So this is kind of the utopian vision of what the Internet of Things could be.
And then another view of the Internet of Things is in terms of the Gartner Hype Cycle. I don't know how many of you are familiar with this, but Gartner puts out a new hype cycle every year in roughly August, and it always has this structure. So across the x-axis is, essentially, time, but instead of being labeled in years or months it's labeled in these different parts of a technology's development. So there's a point where a technology starts to take off. It reaches a point that we call the peak of inflated expectations. So that's like maximum hype for something. And then it drops down into what they call the trough of disillusionment.
So we were so excited about something and then we realized it's going to be harder to than we wanted to. And then we sort of climb back out to a point where we can actually build these things as useful technologies. And so the relevant thing here is that when I arrived in DC, the Internet of Things was at peak hype, and there were other IoT-related things sort of coming up the pike behind it. And if you look at the more recent Gartner Hype Cycle curves, you'll see different technologies. Blockchain and so forth have sort of occupied the peak hype instead. And my observation is that, to some degree, the government policy people, not just our own but across the world, are sort of really agitated about whatever is up there around peak hype and trying to understand what the implications are.
So today's IoT has this kind of consumer-facing view with forks that tell you when you've eaten too much or a thermostat that figures out what temperature you like it. But there's also maybe more satisfying, inspiring ideas. So helping with agriculture, helping with environmental sensing, medical devices that can help someone survive cardiac arrhythmia or seizures better, and also the more infrastructural and manufacturing sides of IoT. And it turns out that the potential for economic growth, first of all, is huge.
But second of all, is sort of split roughly evenly between this outward consumer-facing stuff that sometimes gets joked about and the more sort of serious infrastructural stuff, like the electric grid and so forth, that could be very influential and also very helpful. The other thing that's interesting about this mix of devices and markets is that you can see, definitely, why policy might be relevant. Because maybe you don't need too much policy for an IoT-enabled fork, but if it's going to be something to do with our electric grid or a medical device, it starts to become relevant for policymakers to be involved. And so you can sort of go through. Policymakers worry about things like health, safety, the environment. Policymakers worry about implanted devices inside someone's brain. I have benign hand tremors, and some people were saying that implanted devices can help reduce tremors both for me and for more serious things like Parkinson's. So there's a range of those kinds of issues.
Security comes up in the IoT, as well. I've given this talk periodically over the past few months. I haven't updated it. I could update the slide every couple of months with the latest security hack related to IoT issues. So this is a big, and I would say looming-- we need to do better on this. And then there's privacy in the IoT, as well. So the notion that we have these data-gathering devices all around us. How do we manage our own data and our own portrayal of ourselves in the world? So how do policymakers get involved in IoT? I would say that there's these main categories-- so technical standards, consumer protection, infrastructure protection, and trade-- and they get handled in different forums. But one of the things that I worked on a lot were these technical interoperability standards, which are key to IoT because you have so many different heterogeneous devices that are working together. And so if you don't have good standards, they won't work together and they won't be very useful. Radio frequency spectrum is another place where IoT is thought by some to have an important role. How do we make sure we have enough wireless spectrum for all the IoT devices that are expected to be coming on board soon? And is there a need for particular spectrum bands which some countries advocate for? Or should we just sort of allocate within the bands we already have? Consumer protection is obviously huge, both in terms of the reliability of the systems and also in terms of the privacy implications. Critical infrastructure-- the electric grid and transportation is a place where governments are expected to play a role in ensuring that the infrastructure is safe and protected. And actually, one recurring theme was folks wanting to make technical standards about these topics at the ITU, and the US pushing back and saying, we want a safe electric grid as much as anyone but you can't have international technical standards on the electrical grid because that's our sovereign critical infrastructure and we have to decide how to protect it ourselves. So I think the way I would put it is there were issues where I didn't have to take a position on the issue itself, I had to take a position on whether this was the forum to be discussing that issue. Question.
AUDIENCE: I was wondering if these issues that you deal with were informed by academy panels first. [INAUDIBLE]
MARGARET MARTONOSI: So there were a range of ways that the State Department and other folks would get their information. So actually, it was an academy panel that suggested they form the JSF program to get active scientists into the State Department. But in addition, there were tons of panels going on. While I was in DC as a government person, I was also serving as a subject matter expert for an academy panel that was stood up by the government accounting office on a request from Congress to study IoT. The Department of Homeland Security was also doing a report on IoT, and there were private sector nonprofits doing reports on IoT. And also, over the same time period, the Department of Commerce put out a massive request for comments and got back hundreds of responses, both from individual private citizens, from technical experts, and from different non-profits and companies about IoT. So it was like John Deere, IBM, Verizon, and then random citizens. So there was a lot of input. If anything, there might have been a signal to noise ratio issue more than a lack of information issue.
Right, and then the last thing is just there are folks who see IoT as being a chance for an economic leapfrog. So you'd see these dynamics where certain companies or countries were eager to shape the discussions and the policies in a way that would help them. So where do policymakers get involved in IoT? The previously was how did they get involved. When it comes to where, it's like, where don't they? It's sort of all over the place. But my hope was that-- there were a lot of discussions that weren't very technical, but I wanted them to at least not be wrong. So if I was in the room, I wanted it to not be wrong. That was my goal.
I will dig into one experience. I was asked if I was on the microphone. So this November. This was October, November, 2016. So this was at the end of my stint. So I was the spokesperson for all the IoT issues and all the study group 20 issues that came up at this conference. This is the every four year conference that sets the agenda for the ITU-T for the next four years. So if you are a country or a company that wants to use the ITU-T to advance some goals that you have, you would come to this meeting and try to negotiate that into the outcomes of the meeting. So again, ITU is this UN agency on computing and communications.
And then the World Telecommunications Standardization Assembly, or WTSA. It's every four years. It was about 500 people. It was about 100 countries, and it sets the technical agenda for the next four years. So one recurring theme-- which if you've studied internet policy, you've heard of this, but for everyone else-- the internet operates through a model of self-governance that is quite unique. So the internet operates through stakeholders which could be companies, could be nonprofits, but it also could be just interested citizens. So any stakeholder can come to internet governance meetings and help shape the policy of the internet. And that is a model that people feel has worked quite well. It has been very open, very transparent, and it's avoided a fair number of political squabbles by keeping the countries, the governments, out of operating the internet. On the other hand, there are countries that have called for the ITU to play a role in governing the internet so that countries can shape internet policy rather than stakeholders. They tend to be countries that want to control content and connectivity within their borders. And so the official US position has been that there should be no operational role for the ITU in the internet, and the internet shouldn't be governed by countries but by this multi-stakeholder approach where anyone can get on the conference call and speak their mind. And so, basically, whenever something happens in the ITU that involves the internet, people get very edgy because they're worried. Well, US people get very edgy because they're worried that they're about to give away the internet to the ITU. That's, like, the constant state of nervousness. Yeah.
AUDIENCE: You said what the US official position is. It's probably a dumb question but how is that captured?
MARGARET MARTONOSI: I have a slide that will get to some of that, I hope. So the US' agendas are by written contributions that go in. This particular IoT privacy was one, and then these other things were other issues. And the general theme is watching a technical issue by my viewpoint as an engineer get hijacked into something more political than technical.
I'll talk about the process first, and then I swear I'll get to the US position. So most WTSA is attended as government-led delegations, with industry people usually attending with their own country's delegation. So the US had representatives from Apple, Facebook, Cisco, Verizon, on the US delegation there in Tunisia. The countries aggregate into regions. And if there is some written contribution that you want to give as input to the meeting, if you can get your region to coalesce on that input contribution, then it becomes a regional one. Then, it has greater weight because it's a whole set of countries behind it rather than just one. So the input to this meeting are these written contributions, and the output resolutions. And the resolutions guide what this organization is going to do for the next four years. So the whole game for people who want to play the game is to try to get their stuff into a resolution so that there's a mandate to have the ITU focus on it for the next four years.
So what is the US contribution? Before a meeting happens-- Sorry, what is the US position. Before a meeting happens, there are weeks or months of preparatory meetings that the US runs in an open way. You could all get on the email lists. The email lists tell you a phone number. You could call in. They are having prep calls essentially weekly or biweekly for something that is the next meeting. So months in advance the US begins preparing a contribution, and that contribution summarizes the US position on an issue that's important enough to submit something and write about it.
So how do you get to a US position? Someone has to care enough to write something, and then everyone else on the phone call can say, whoa, wait a second. And so at lunch today I was talking about one example which is, not at this meeting but at a radio conference meeting, there was a frequency spectrum issue where US terrestrial spectrum users-- which is to say cell phone companies-- had one position, and US satellite companies had a very different position on how the spectrum bands should get used. So they have to hash things out during the US preparatory process because the US goes into the meeting with only one position.
So the argument about the satellite companies versus the cellular companies has to happen through these prep calls in advance. And it gets articulated as a position. And then once that's decided on, the person on the microphone has to stick to that position. So you can't freestyle and decide that suddenly you want to support the satellite folks when the meeting beforehand did something else. You can make subtle adjustments for strategic reasons, but you can't change the position unilaterally. And that was a little hard and weird to have to stick to a position kind of no matter what.
Right, so you write contributions that convey the positions. After the contribution deadline, you can see what other folks have written. So it could be that another region of the world has written something that we are not happy about. You can write a position paper, which is essentially a bunch of bullet points for what you're going to say at the microphone when that gets discussed. So contributions get introduced. Some of them are pretty boring about structural stuff. Some of them are about mandate. And there is this very disciplined set of discussions about queuing for the microphone, and so forth, as I mentioned. There's also smaller ad hoc groups where there's maybe 20 people in the room. It's more free form, but you still have to queue for the floor.
There is very little of it that is technical here. You're literally negotiating text using Microsoft Word with track changes turned on. So, privacy. Privacy is a sensitive issue for a lot of us. The US position here was that we didn't want privacy to be something that was internationally standardized. We could have our own domestic discussion about what privacy should be afforded to the people in our country, but the idea was that it shouldn't be internationally standardized. The one international agreement that the US was willing to sort of sign onto repeatedly is that, for any country, one's privacy online should match the privacy that that country offers you offline.
So where other parts of the world, like the EU, consider privacy to be a basic human right-- it's a civil right-- the US hasn't written privacy into its civil rights in the same level of formality. But it does say privacy online, on the internet, should match privacy offline. There were folks who wanted to have privacy standards be a part of the technical standards that got discussed at these meetings. And the concern was that that would widen the mandate for the ITU from sort of technical standards to more of a human rights group in a way that didn't seem appropriate to the mandate of the group.
So this is usually where someone says, I don't ever want to be a Jefferson Science fellow. But it was sort of fascinating to watch these things play out and to see why people were pushing. And in many cases, they were pushing-- I'm sorry but I hope this doesn't offend anyone, but when Saudi Arabia is the one advocating for privacy to be part of the study group, I don't believe that it's really about privacy. I think it's about they wanted greater technical power within a study group, and they wanted to expand that study group's mandate. I don't think that Saudi Arabia plus its allies in Egypt, UAE, and Russia were really the ones advocating for my privacy. So one of the nice things about giving this talk now is I used to give this talk as a government employee and I had to be quieter.
Another example that came up was about technical neutrality. There were a bunch of contributions that came in advocating particular technology from a particular-- essentially a company, basically a nonprofit. And the US plus like-minded countries were arguing that this was essentially a commercial product placement in the middle of a UN resolution. And that was a whole eight day discussion that went down to the wire, actually the closing plenary. So this is a meeting where they tried to sort of give a compromise text. The US and allies, basically the whole Americas region, refused the compromise text.
And so at the closing plenary, we're running about six hours late. The foreign minister of Tunisia is there for the gala closing, but we have not agreed on that wording that's highlighted in yellow. And it seems to be a complete stand off. They call for a coffee break. So before the coffee break, the standoff is basically they keep trying to have wording that's going to go across these six resolutions, and the US, my supervisor, who had an amazingly calm demeanor, she just kept getting on the microphone and saying, thank you very much but I don't think that will work for us. And she did that over and over again with additional support for why it was technically inaccurate to have the same text on six very different resolutions unless it was about a commercial product placement kind of thing, trying to get a mandate for that company's work to be in UN agencies.
So they call for a coffee break. We're hours past when we're supposed to adjourn. And the coffee break is a whole bunch of people talking. And there's a point where I, with my shaking hands but my camera on massive zoom, took this picture. That's my supervisor, the US head of delegation, talking with the head of delegation from the United Arab Emirates, and he's giving her compromise text. And moments after this happened, Russia came to her with compromise text that was not nearly as good, and she tells Russia, I'm sorry but I think I have what I need. They come out of coffee break, and I don't know if you can tell but there's a whole lot less yellow text highlighted. So that's the text that UAE agreed to. And while this is happening, the countries that had been in an alliance with Russia are literally running down the aisle to talk to Russia because Russia has been sold up the river without their knowledge, and basically get Russia to not object to this wording.
AUDIENCE: What does that say?
MARGARET MARTONOSI: It's wording that basically says that a particular technology has been shown to be useful in one case that may not be relevant in the case of those six resolutions, and it's worrying that's going to go into the minutes of the meeting rather than into any resolution. So basically, people wanted advocacy for this company in a whole bunch of resolutions to create a mandate for it.
AUDIENCE: What kind of technology?
MARGARET MARTONOSI: Just a company. Imagine if we had a commercial for Netflix in the middle of a UN resolution. It was something at that level.
AUDIENCE: But I mean, you're not saying specifics so it's really hard for me to--
MARGARET MARTONOSI: Digital object architecture.
AUDIENCE: I mean, when you say company-- what, like robotics?
MARGARET MARTONOSI: Digital object architecture.
AUDIENCE: Excuse me?
MARGARET MARTONOSI: Digital object architecture was the name of the technology. Digital object architecture. So it's not a widely known technology. They were advocating for it to try to get more widely known.
AUDIENCE: Right, but what is that?
MARGARET MARTONOSI: That's a whole other talk. OK, so in summary--
There are a whole lot of these forums going on. Sometimes the technical topics do get discussed deeply. Sometimes they get brought up as a vehicle for something else. And I think, sadly, sometimes what makes sense technically doesn't carry the day. But even so, even with the 1:00 AM footnote discussions, I still see a lot of value in educating policymakers about the technical details behind all of this. I have after April gone to talk to my representatives, tell one senator, not my own, about how blockchain works, and sort of make sure that there are people within our government who understand some of the basics of the technology so that they can understand where the hype ends and the technology begins. And it also motivated me to come back to my lab and work on more deeply technical solutions for technical problems, like IoT privacy, so that policy and regulatory solutions weren't the only option. So if we come back to here, that meeting that I showed you all the pictures of ended on November 3. You can do the math and think about what happened a few days later. So, you know, this is ongoing, right? Every time I give this talk it's slightly different.
The new head of CIP was just named and came into the office about two weeks ago. He was a former tech staffer in Senator Corker's office. Senator Corker has been in the news lately, so he found himself a new job pretty quickly. And you'll also notice that his official title, where my office when I was there was named International Communications and Information Policy, his official title has added Cyber back in. So it is this merger that I spoke of between the SCCI and the CIP.
OK, so what are ways to be involved? I think the general advice that I have is to find people who are open to hearing more about technology, and try to explain it to them and have influence. This is timely. It didn't happen yesterday, it happened three days ago, but I just found it hilarious. First, I read the top thing and was like, oh, that's pretty cool. This is on a Facebook group that I belong to that's about political advocacy in New Jersey. So that's kind of cool, net neutrality and stuff. And then I'm like, wow, that is a unique way to have influence on the technical landscape. So I'm involved in the top paragraph. I have not volunteered to be part of a theater troupe, and I have no idea what the [INAUDIBLE] thing would be about.
More seriously, there are a bunch of ways to be involved. I mentioned that the Commerce Department, NTIA, did a request for comments on IoT about a year ago. And the FCC is often doing requests for comments, including an ongoing-- the FCC requests for comments about network neutrality has received more comments than any other RFC, I think, in government history. And what I can say is that when they do these RFCs, there's a range of responses. So some are the public policy offices of a big company. They kind of say what you expect them to say for that company's position. Some of them are tinfoil hat people who can write some pretty funny stuff. And then some of them are knowledgeable, well-reasoned responses. And I know that, for example, one of my colleagues wrote a very directed piece in response to FCC, RFC. Thought he was just sort of sending it over the fence and would never hear again. The FCC called him a week later to say, we want you to come in, we want you to teach us more about your views on how measurement and research can be influenced by what we're doing here with this policy. Congressional representatives-- it's not just federal. A lot of this is happening now at the local, and state, and regional levels, and it's important to sort of contribute to that, as well. And it's not just government. There's also these nonprofits and, like I said, ways to engage as a citizen multi-stakeholder, as well.
So why should you do it? If you're a scientist or engineer, you have deeper knowledge and technical expertise than most of the other people that are talking to our policymakers today. If you are an academic, I have nothing against companies, but I will say that companies have huge public policy staffs on their salary to advocate for them, and academics are not on the radar screen to the same degree. So if you want a broader and more neutral perspective to be part of what gets heard then you need to be one of the people who says it. And then as a person, because you can, and I can assure you it will help.
So I will end with this picture. That is from Tunisia. That is a sunrise. So it's not a sunset. This will continue, and thank you very much.
AUDIENCE: Could you comment just briefly on your impressions of this whole process from the female scientist's perspective? Do you think that played any role?
MARGARET MARTONOSI: That's a great question. So from a female scientist perspective. So I will say that I'm a female computer scientist. So going to DC and working in an office in DC, I was for the first time in my life in a majority female professional environment. And it was also an older environment than I expected. DC, the legislative side, has a lot of the 20-something staffers, and it's a very young city in many ways. But I think in the civil service parts of DC and in the executive branch agencies, you have career people who span a wide range of ages, wider than the typical academic or sort of techie environment. So I both felt sort of more in the middle in terms of age and I did not feel like a minority in DC. When you go to these meetings, it gets a little weird because there are a lot of cultures that are mixing. And obviously, if you're negotiating with a country that wouldn't let you drive until two weeks ago, you know, that's different. But in general, the gender thing was not as overt of an issue as you might think. Anything else? Yeah, back there.
AUDIENCE: What is your opinion on the recent stories about scientists and engineers trying to actually run for government? Do you think they can be effective, or is doing stuff like this more effective than pushing policy?
MARGARET MARTONOSI: I think people have to make their own choices. I think I would love to see more scientists in government. I don't know if anyone was here when I--
So Rush Holt is a PhD physicist who was Princeton's representative in Congress for close to 20 years. He's now CEO of the AAAS. So he's sort of considered like a big congressional scientist, a leader for many years. And there was a point in January where I emailed him on a whim and I said, how does one run for Congress? Because I don't live in the district where Princeton is, which is deeply Democratic, I live one district north. And I was surprised. So I emailed him just on a whim-- not a whim, a principled whim, we'll say-- and he replied that night and he said, here's my cell phone. Call me. I said, I'm not ready to call you. Give me a week to get my thoughts put together. So I sort of sorted through this. It takes millions of dollars to run for Congress right now, and it takes a leave of absence. So it's not clear that that's for everyone all the time, and that's why I think there's a range of things to be done. I would love for there to be more quantitative and technically comfortable people in Congress. I think that would change a lot of discussions. I would love for there to be more women in Congress. I think that would change discussions, too. But I think there's ways to have impact without running for Congress, as well, and I'm sure you could find me-- you know, it's not like all scientists are or good, right? I'm sure you could find me a scientist that would frustrate me, as well.
AUDIENCE: Yeah, I would like to go back to one of the things that I was a little frustrated with your talk is you didn't give a lot of detail. So you're talking about more of a skeleton example of what you did, but I don't know what kind of policy you were working on. A lot of the issues I'm concerned about is we, as a human being, my rights being infringed by technology, you know? Like I mentioned before the lecture started, all these people that are part of the US Facebook, slogan and we're learning now that these companies are manufacturing devices that keep us occupied to them. And I don't understand. Like, OK, so your job is to help the government do what? Like, what is your--
MARGARET MARTONOSI: These were technical standards meetings. So they were writing technical standards for how technologies interoperate, and in particular, one whole study group was about how Internet of Things technologies might interoperate. So there are communication protocols, different aspects of how the data was collected, and so forth. And our hope was to limit the degree to which countries, governments, would be involved in articulating anything about how the Internet of Things would collect data because most of us would prefer for the government to not be involved in those kinds of technologies.
AUDIENCE: I mean, these companies, the Internet of Things-- I mean, that's such a broad thing you're talking about. Like, you know, there's issues already with cars that drive themselves. I watch a lot of news, and I'm politicially if I could say that. And already, the technology is moving faster than laws can be applied to them. And that's a very serious thing.
MARGARET MARTONOSI: I totally agree.
AUDIENCE: And you know, we're concerned about climate change, or what cars are doing now, and older technologies, but we're not talking about what the new technologies are actually doing.
MARGARET MARTONOSI: I disagree. I think we are talking about it a lot. And I think there is a place for domestic regulation of these issues, and I would love to see it. I think there is a place for domestic regulation of a lot of issues. I'd love to see the EPA regulate some things, too.
AUDIENCE: But you in your position, I mean, that's what I'm getting at. That's why I came here, but you in your position, have you advocated for that?
MARGARET MARTONOSI: Domestically, yes.
AUDIENCE: OK, that's what I wanted to know.
MARGARET MARTONOSI: I advocate for it based on my own opinions. You can advocate, too. The point is the process is there.
SPEAKER 1: Well, we are a little bit over time, but I am very grateful for everyone that came. It was very fascinating, so thanks to Professor Martonosi for sharing her experience. She'll be here until Friday, and tomorrow we'll have a more technical talk on the research that you're welcome to attend, as well.
We've received your request
You will be notified by email when the transcript and captions are available. The process may take up to 5 business days. Please contact firstname.lastname@example.org if you have any questions about this request.
We live in a world where computing devices and systems have more societal impact than ever before, and yet they are simultaneously more difficult to test, debug, patch or even understand than ever before. Questions are constantly arising regarding how reliable and secure they are, and some autonomous systems are even raising broader discussions about fairness and ethics as well. In the midst of this, policymakers are—often with relatively little technical background—deciding what to do in terms of regulation and management of these systems. While computer science plays a prominent role in policy circles, policymaking for other science and technology fields like climate science, medicine, and many others is also both very complex and deeply important.
From August 2015 to April 2017, A.D. White Professor-at-Large Margaret Martonosi was a Jefferson Science Fellow (JSF) engaged in computing and communications policy within the United States Department of State. Established in 2003, the JSF program is one of many governmental programs aimed at improving and augmenting in-house science and technology expertise.
In this talk, given Oct. 11, 2017, Martonosi discussed the role of scientists and technologists in policy circles, drawing from her own personal experiences. She concluded with some thoughts on how—and why—scientists can and should be personally engaged in these issues over the months and years to come.