stanford stream of spoly institute for the national studies chair of stafford's artificial intelligence and international security steering committee and a contributing writer at the atlantic but wait, there's more abc a guards also a book author her latest title being spies lives and algorithms the history and future of american intelligence. amy great to see you and congratulations on the book great to see youtube bill. thanks. it's always nice to your book rather than just be writing a book and that leads to my first question. i just listed countless things you're doing in and around the stanford world. how do you find time to write a book? well because we couldn't get together and we couldn't travel so i am moved all my books to my home office and hold up and that's how i was able to. finish it took a lot of you know structured time away from the office to get it done and just a lot of discipline, i guess and just kind of keeping atticus. it's like i deal in collaborating but i'm writing and you know 800 word bursts here and there but a book is a is a different creature. well, i can't do what you do. i i'm afraid of writing on deadline to be honest. so if i have a deadline that several years away i can get it done. but if it's you know several hours away, that's another story stick to what you do. what you do is thoughtful you're not living in the moment. so well done. so let's talk about the book spas lies algorithms. let's break it down into categories. amy spies. what are we talking about? we talk spies these days. well the let me take a step back because the purpose of the book really is to inform a general audience about what the secret world of intelligence is spies usually get people's attention and i think there are a lot of myths about what spine is even the basic terminology we call someone a spy when they're not really a spy. the spy is the foreigner that bet. their country for our cause right not the cia officer that actually is running them. i think you know that one of the main points of writing the book is to dispel some myths right, so i'll give you like my top three myths of good community myth number one, is that intelligence is secrets, right? most of intelligence actually is not secrets, right 80% of a typical report is open source information, right? myth number two intelligence is policy, right? that spies are out there giving policy advice, right mr. president. you should do x y or z. they don't they're not supposed to do that. they usually do and i think the third, you know sort of big myth is covert action is this bag of dirty tricks that we reserve for the you know, the most horrible things that our government does and that's not true either we do everything overtly that we do covertly what makes something covert. is that the us government tries to it's official responsibility for it. so who's our covert action wars are over action to replace regime. so we do everything overtly that we do covertly right so it comes to my name is this documentary? it's up right now and oliver stone in the movie jfk which justice celebrity anniversary and it gets the two things number one how stone made the movie but secondly, it's oliver stone opening his head and revealing what he thinks about the jfk assassination and vietnam and intelligence. so before you know, you're on a very dark road or what the us government does and doesn't gun and now we're living the days of the cia and cuba and trying to feed, you know fidel castro poison cigars and things like that. and so that's what you're getting out there. just the public is being fed a lot of information. that's maybe not quite accurate. yeah, and what we see bill is you know in in public opinion polling is this dramatically rising belief conspiracy theories of all types. so one of the most stunning polling results i found was that even if you years ago, 25% of americans believe 9/11 was an inside job by the us government a quarter of americans believe that so a lot of what i'm trying to do is dispel these, you know crazy conspiracy notions about what the intelligence community is really doing right, but that also leads in a technology. we'll get that no second. i mean you take that poll of a quarter of american snake that 9/11 was an inside job. i think the actor charlie sheen was a big proponent of that theory. i'm not sure what paul say about people thinking about vaccines these days, but i'm sure there's a healthy portion of population who thinks of vaccines. is this for one thing that's allowed bill gates to microchip you so, you know, so i don't know if it's technology is driving this per se or maybe it's the medium pop culture amy, but this is what i think you're getting at at your book. it's just it's a lot of a lot of what the government does and doesn't do on intelligence is sort of polluted by what pop culture gives us what academia teaches us. yes, so what it's it's i'm glad you raised that bill because what originally got me interested in writing the book which was many years ago when i really started thinking about it was a poll. i did of my college students at the time at ucla and on a lark. i just asked them a bunch of questions about intelligence and then what their television and movie viewing habits were and what i found was that a statistically significant percentage of them? were affected by their spy themed entertainment or at least there were correlations those who said they always watch the show 24 for example with jack bauer were far more willing to advocate really aggressive intelligence policies like waterboarding for example, right? and so what i found the more i dug into this was that spy themed entertainment had actually become adult education and i found all sorts of evidence about this and national polls and actually in the policy world. okay, so the book you talk about espionage going back to the days of george washington. so it seems as government has shall i say a rich tradition of being involved in espionage, but very simple question amy, how does espionage differ today versus it did back in the days the founding fathers? oh, that's such a good question in some ways. it's really similar more similar than we might think so. we think about information warfare as an internet invention, but in fact benjamin franklin was really good at information warfare cranked out fake news literally fake news articles from his paris basement. it's different today. primarily. i would say in three ways. number one is speed. everything's moving faster. now the speed of data the speed of insight the speed of decision. so espionage has to keep pace with what decision makers need to know and when they need to know it, so everything is accelerating today. that's a really hard challenge. i think for intelligence agencies, right? the second is scale. so if we think about i have a chapter in the book about traders and counterintelligence, you know, it used to take years for people to smuggle documents out in their pants and garbage bags and all sorts of crazy ways to to try to betray their country, right but now traders can download documents millions of documents in a matter of minutes or hours or months. so this scale of espionage and particularly counterintelligence challenges is completely different and i see the third key difference with espionage today is that there's been a democratization of capability. it used to be that that superpowers like the united states had a really advantage in espionage right only the government, really and maybe the soviet union could launch billion dollar satellites and have massive capabilities of code breaking and code making well now anyone can gather that kind of data and anyone can analyze that kind of data ai capabilities are available on the internet. they don't require a degree in computer science. we have satellite imagery. we can all you know, if you have an act if you have access to the internet you have access to google earth, so there are and and events are live tweeted right so we can track things on twitter. so now we have to think about intelligence competitors today being much more spread out and it's a much more crowded playing field for us intelligence agencies. okay. so who are the competitors? well, the main competitors are of course nation states, but anybody can do this now. so if we think about on the good guy side of things right and think about nuclear threats, there are all sorts of people outside the us government that are tracking nuclear threats and they're doing a really good job. some of them are colleagues at stanford, right, but you know in the in the past year for example news came to light about these hundreds of chinese nuclear missile silos that were previously unknown. well that came to life because of people without security clearances without access to classified information just using commercial satellite imagery and their expertise and posting it online. so people who look at google earth shots of chinese naval yards and north korean missile sites and whatnot and then say look something looks bad. but why are they doing this amy? why isn't the us government or is the us government doing this? we just don't know about it. you mean to do is government tracking these things or is the us government working with these people right? a little scary to think that some individ. not a vigilante but individual amy is looking at google earth and then reporting that look at these chinese ships being built in the naval yard, and the thought is does us government notice on i'd like to think the us government has step ahead of that of that individual. so the us government is aware of what's going on and i would say and in many cases partnering with non-governmental nuclear sleuths as i roll them, right? so there are partnerships some are more formalized and others, but you put your finger on a key point bill, which is that right now this ecosystem at least in nuclear security is dominated by americans and our western allies. that will not be the case in the future right? it's open anyone can join this world, and we've already had instances where so we say less benign actors or nefarious actors are deliberately trying to inject falsehoods into this ecosystem in the hosts at the us government will fall for it. so it's going to get more complicated will be bad actors and or of them in this ecosystem in the future and the government's aware of it and trying to figure out ways to engage more productively with this non-governmental ecosystem. right? let's talk about the us government intelligence apparatus amy. i was watching the army navy football game this weekend and during the third quarter commercial came up in the commercial was trying to get you to join the army to get involved in counterintelligence to stop hackers and then it shown on the tagline. it showed all the branches of the armed services including space force. that caught me to thinking who drives the train these days. is it is it the military is it the nsa is that the cia is in honolin security we seem to have a lot of entities devoted these days to the great umbrella of intelligence and counterintelligence. well, the unsatisfying answer who drives the train today. is it depends? okay, so we have 18 different agencies in the us intelligence community today that number often stuns people 18 agencies. we hear about the cia maybe the nsa but they're a lots right and that's much higher than it was 10 years ago or 20 years ago. so whenever we have a crisis the tendency is let's create a new agency for that and the result is it's really hard to coordinate them all so nominally in charge of this behemoth community is the director of national intelligence that's getting better, but the dni has only limited control over budgets and people and as you know in washington, those are two very powerful lovers if you can't control the budget and you can't control the people you're in the business of trying to persuade and joel right not direct and command and so it's a work in progress how this community works together. and do you think the 18 need to be folded into a dozen or six or would you would you downsize that or would you would you create a super agency to walk over 18? well, would you streamline this? well streamlining is hard, but you know, i people arm wrestle over this all the time, right or do we have too many do we not have enough the benefit i give you the the argument on one side, which is that the reason we have so many agencies is they provide tailored capabilities, right? so the navy has different intelligence needs than the army and so the navy should have its own intelligence unit to help it with its intelligence priorities there's real truth to that and different agencies have different specialties. so cia is human intelligence, right national security agency is signals intelligence. so email phone calls, right and specialization is has benefits. so i always give the example of doctors, right? no one questions whether doctors should be specialized. you don't want your heart surgeon detecting skin cancer. you don't want your dermatologist operating on your heart. so specialization has benefits the challenge though is how do you harness that specialization? so, you know what everyone knows that's challenge and that's where emboldening and powering the dni which has happened over time has paid some real dividends, but it's still it's still a challenge. it is a challenge maybe getting the agencies to talk to each other. it wasn't this one of the takeaways from 9/11 that digest the is just not a cross-chatter with them washington about what intelligence we had. yes, and that cross chatter communication is much better than it used to be which is not saying a whole lot right because the bar was low before nine eleven, right, but it is getting better. i have a heretical idea which i've had a lot of people's or arguing with me about which is what we actually really need most is a 19th intelligence agency and i say that with some trepidation because of the coordination challenges we just talked about right the 19th agency would be dedicated entirely to open source intelligence stuff that's publicly available out there on the internet because what's happened in our in with this american, you know, we're talking. out technology because of the advent of new technologies right the whole intelligence battleground has changed right? it used to be that secrets were more of a key secret still matter, but now insights coming from harnessing lots of openly available data and secrets will always be primary in secret agencies. so no no existing agency is giving public information the attention that it deserves and you're not going to get that unless you get a new agency trying to harness the internet to be saying it sounds a little bit like drinking from a fire hydrant in terms of collecting information. so how would you actually physically what you just have an army of pardon me nerds and pajamas and basements around world going through going through and how would you how would you actually attack the unit at that way? well, a lot of it can be you know, how can technology help us with technology? yes, how can things like ai tools augment the human analysts, right? we have just far too much data for anybody to process at any one time. to give you some idea, you know the amount of data is estimated on earth to double every 24 months, right? that's an astounding level of data. but algorithms can help make sense of that data, right? so i'll give you a concrete example that happened actually at stanford within the past year. so i have two colleagues who wanted to better understand trade between north korea and china. and so what they decided you was look at the imagery of trucks crossing between the border between these two countries and let's go back several years. they thought and let's analyze the truck traffic between these just to get a sense of what can we derive from looking at this and so without any computer science training, they developed a machine learning algorithm so that the machine would automate the scanning of the trucks across the border, right? and what would take a human analyst roughly a month to do they're very basic algorithm did in 20 minutes. that's the kind of benefit we can get from harnessing technology to understand this overwhelming question data. right, but you would still need the human element to eliminate what a certain somebody would call fake news. absolutely, and when the idea is that you if you you humans can do best no machine is going to be able to define the intentions of north korea for example, or to consider alternative hypotheses, or to ask. why is this the case right? it can just have pattern pattern recognition is the real benefit of these tools. okay, so where would that 19th agency amy said in a flow chart of the federal government. you're really -- me and a good way bill. so i would say the 19th agency would be independent. there's been a big debate about could you put such an agency in the cia? could you put it in the state department? i think it needs to be a standalone agency. and here's the thing. it should not be inside the beltway or at least not entirely inside the beltway because if you want to attract the best minds of tomorrow, you need to go where the talent lives right? so imagine a forward. floyd agency which with offices in places like silicon valley and austin, texas and denver, colorado, and now you've got this open source stuff so you can experiment with new technology tools in a more ongoing way. so it's not just the stuff that an open source agency could provide it's the people and it's the processes of innovation. okay, see what an agency that's independent and outside of washington. yes. you are a heretic. i am indeed. it's it's time for heretical thinking though. okay. i want to read something you wrote for political back in september amy's. this is my way of saying good time to grab the water if you've got it sitting by you. so here's what you wrote. this is on the 20th anniversary of the 9/11 attacks quote 20 years after 9/11 the united states faces escalating threats from china russia, iran and north korea conflict in cyberspace as well as physical space and global challenges like climate change and pandemics the cia needs to regain the balance between fighting and terrorist enemies of today and providing the intelligence to detect understand and stop the enemies of tomorrow. yes, and you know, this is something where people inside the intelligence community have been discussing this balance for a long time. right and it's interesting bill. i got a lot of feedback from that article more than i have and just about anything i've written in the past couple of years people inside the government saying that they thought this was exactly the case. so in intelligence, you have to balance between dealing with urgent and dealing with important threats, right? my argument was there's no organization other than the cia. i mean whose primary mission primary mission is preventing strategic surprise, right? and the more the cia gets sucked into the day-to-day counters counterterrorism paramilitary activities and supporting them the less time it has to prevent strategic surprise with all these other threats in the world. and so the balance has tilted far too much toward tactical warfighting intelligence. and that's not say it's it's unimportant, but it has to be in balance and the cia's balance is out of whack and it needs to get back to a more balanced portfolio of activities the last time the cia's mission would have been examined. amy would have been what the mid 70s. i think the cia's mission was examined pretty carefully after 9/11 and the surge right of activity and we got pretty good at the sort of support of counterterrorism activities. so it's a it was