Deepfakes and Fake News

McAfee as a senior sales consultant please welcome to go start with the disclaimer this favor is that the public informations and there be containing strong language that even if they would never save those states so for instance they could have me say things like no kill monger with white Ben Carson is in the centum place or simply President Trump is a total and complete [ __ ] now you see I would never say these things at least not in the public address but someone else would someone like Jordan Peele this is a dangerous time moving forward we need to be more vigilant with what we trust from the internet it's a time when we need to

rely on trusted news sources they sound basic but how do we move forward age of information is gonna be the difference between whether we survive or whether we become some kind of [ __ ] up discover thank you stable [ __ ] so this is kind of interesting right it's a little bit scary when you think about it because you can take almost anybody you make them sound and make them say whatever you want and that's the whole concept of be fake so we will explore that okay look at what the fate is what fake news are and are there any ways today to detect this is there any kind of technology that can be used to detect that and also

what can you do about it and the conclusion of the annual you'll notice that there's a lot of things that were relied on you unfortunately still today so you know it used to cost lots of money to swap a fix now you can get it on snapchat half now and change your face and put funny ears and all that so something will recognize your face and do different things but you know you look for example at Star Wars where this three-second video of Princess Leia 40 40 years later looks like her but that three seconds videos of months to create it cut it costed a lot of money same thing for kernel target okay commander Tarkin same thing Peter

crushes peter cushing that died in 1994 reappeared even the latest Star Wars as Peter Cushman and it was essentially an actor which as I think is his son that they took to make him look like this so there aren't possibilities today to recreate personas digitally to make it them quite good it's still you know with waxey is still a few things that are kind of you know not exactly natural but it's getting there however since then there's even newer technology that is free they will allow you to do this and it was called the fake the actual app was called fake act so there is a fake app pick up the org it's still available but be very

careful because as a krypter krypter minor malware in it so the app works if you allow your machine to use as a crypto minor or coin collection which is you know kind of a downside to it but you can download this app and the requirement to make that app run is very little you need 8 gig around any kind of secret modern CPU and it just need kind of an Nvidia with a cool image in it if you have that yeah everything all Windows 10 and if you have that you have everything that you can that you need to create a deep fake so what is a big fake it's a realistic looking video that will

show event that never added so you can do all bunch of things like this so you can swap faces typically right now it's swapping faces but you can also merge audio so there are techniques I'll show you some of the things after where you can actually use audio to fake argue of somebody I will use my own voice and ask the computer to tell something that I've never said myself ok so it's a priest under free software runs a standard PC and however you do need a lot of images to make this happening to be good and natural so if you have two persons that you want to swap faces with and that's a new video not just an image just a

picture but it being something that moves you need anywhere between 300 and 2,000 pictures of those individuals of each in order to have a good match between the two so that works really well with you know celebrities you know government politics politicians and those type of things it works for you well then however for the simple of the mortals a little bit more difficult because you really have two thousand picture that a girlfriend or a boyfriend somewhere you know that's and that you want to use to make some very very something a little bit more difficult so how does it work its use first kind of be compressed methods so you need to use a compressor so you use kind of an

encoder when you take the original picture of a person and you encode it and kind of shrink it okay wimpy encoder I'm here so you do that in this case what you wanted to do is change a Jimmy Fallon with John Oliver so switch the faces up both in order to do that in the video okay so first thing encode the same person's you encode the person a and making sure that on the other side when you've decoded when you become press and they to be information out and it comes the same way second step it will take a warped version of the same picture so you take a white person in the same picture run into the encoder compare to

the first one and make sure it comes back again like the same person so essentially remove to work so Ken the machine learning because that's going to be neural network can the machine learning understand what the facial is and reconstruct even if it's kind of worked do the same thing with the second persona running to the same encoder and yet you know in that case of our version of Oliver and get this this you know good version of Oliver and at the end you take a good version of the first guy so fallen and you transform it into the second one and you end up with something like this okay so I won't run the video

on this one but essentially you have the face of John over on top of Jim Popham and that is done in a few hours so when you start running this takes a few hours and you create that first version okay so is it being used for good or evil or just for fun there's a guy on the internet decided that you know what I really like Nicolas Cage they took every single character of the Lord of the Ring and replace them with Nicolas Cage so here you have name escapes me go let's get it let's get it goes everybody is Nicolas Cage it is actually so it could be fun could be interesting you know we

can use it for fun but it can be used for other situation as well so here's an example this one is actually pretty cool this is there's a channel on YouTube called dirt face and you have a bunch of good and not-so-good examples and I'll put again the microphone is here in this case it's a it's a James Bond they do and I'm not sure if it's Goldfinger or dr. no but the took Sean Connery and replaced him with Elon Musk the field you'd be surprised pretty cool very simple just need to run a few samples and you get this kind of result with that very simple fake app software again you can do that at home

this one required many more hours from the actual maker of it because it's actually pretty smooth so it's very convincing so if you have somebody that looks a bit like the original actor so in that case it was Elon Musk and Sean Connery and you kind of have a similar fascia so when you replace it this way it looks very very smooth when you do a first pass in a very short period you can know this kind of discoloration around the face and two other things the depth you know right away it's kind of a deep fake but can also be used for evil and when was the first place that this was used on they kind of a white scheme

horn okay so it was done in such a way that said waha you know is there any porn from Scarlett Johansson or Maisie Williams or Emma Watson or you know are just so of course lots of people decided to create porn videos with regular actresses so obviously nobody was really happy about that and she's starting to ban them so there was a lot of banging so it'll read it kind of you know kind of prohibit this a bunch of site pornhub and and and other kind of horns I think this actually prohibited the the usage of fake personas and you say well those are just actresses do we really care well it was also used in many cases of

revenge porn where some people took basically their girlfriend and say Hank she did the point she didn't read but you know the technology is strong enough that is actually good used to be good use for that so it can be use for evil as well so we technologies cool you can do something fun but it can also do something with you quite dangerous the company montreal called Lyrebird ly are eBird created something pretty cool and what they did is that they created a way to create for in avatar of your voice with only reading 30 sentences that's all we have to do you read 30 sentences and you can type now in a box anything and it will do it with your

voice so right now the technology is limited for your own voice they have packages where you can actually kind of by other people voices for automated systems phone systems and others but this by the way is free so you can go to Lyrebird ai okay and to create an account take the ten minutes it takes to create your basic voice and then the replay anything you want I'm gonna show you two samples because I've done it for my own voice it is a liberal body so it's not perfect so if I spend two hours three hours doing the same thing the most would be much more precise but I wanted to meet the judges I can't say it

body but I'd like you to tell me if it sounds pretty close at my boss because we're never good an earring our own voice right it's all straight up

welcome to be signed in for every day I hope you have a very pleasant show so far and that the present er is doing an extraordinary job so that's why it will be sunny today in Fredericton if you can't take a break and enjoy the nice weather so like I said it's so good for body but it's actually pretty close I think to the tone of my voice and if I would practice more than just in ten minutes I can literally recreate my original voice okay so you do have examples of other ones that of course they have Obama they have mr. Trump yeah other ones that he actually created and they are pretty on the spot so this is a free

software this is kind of what's what's coming so think about this you know right now it's locked in such a way that you cannot use anybody else's voice which is a good thing who knows at some point those things could get out and you'll be able to take samples of other people and type and make them say whatever you want so that is kind of a danger of the technology but that's out there as well so how to detect deep fix well some of them are really really obvious so some of them are really easy to spot some other not so much you know like the Elan musk is actually pretty good the first one Obama was pretty damn

good especially you with the voice matching so you have an actor on one side speaking and the no real loop on the other side moves and is it sound exactly like Obama on the other side so you can have a person like this say anything so imagine the danger with mr. Trump down south that decides that somebody creates a big fake of him and say we have just launched a nuclear attack what could be the implication - that will be pretty bad so how do we you think those there are different ways to detect some of the information some of them like I said are easy to spot some of there are more difficult to spot but when you

record while you take a picture or something either video wise or a camera a photo there's always kind of an imprint on it so we have you know in the photo you have the EXIF information or the XMP information and that will tell you what camera was being used yep it sure was using the speed and how much of other information is regarding that specific video or image this is actually being you can actually see this and if somebody tampering information they can be a gleam in the actual video where that information disappears or is replaced or is changed if they do splice or it isn't somehow you can recognize that but like anything else it could be

recreated to make it seamless so that information is being kept so what are other techniques that can be used so there are some machine learning right now techniques that are being developed then though by DoD specifically to try to find out if somebody fakes something or not and you can look at something like linking of the ice if the blinking is not natural and not up during a certain interval it will say there's a hint looking at a pulse in the neck so if you can detect a false and it's regular then you can say okay well that's probably true if there's no pulse or the pulse is not regular that something has been altered somewhere

along the way so some of those are some of the techniques that are actually being used or are basically being trained today to try to detect those things okay however there are no commercial software right now to do this so this is all still on a development basis let's talk to Peggy's I think it was a great thing right so if we look at the US president is everything that's negative about it that's big news that's very simple for everybody else it's something misleading or false and there's a big human factor in there the human factor is we sometimes even agree if something is good or not you know this pretty good example here how do you know this isn't

this news story is true well if I agree with it it must be true okay so that there are ways of how do you distinguish this this is why it's so difficult it's not just coming cast and concrete so it's black white white you know there's something in between is it really true is it's not true where is it and that is even the much more difficult area right now to tackle so this is that new fake news have been there for a long time and again how do you define fake news it's how you there's something misleading it's maybe a falsehood something like this so we'll take the example of the snake oil bender way back

then that existed they also had other campaigns where they were modifying pictures of presidents you know to represent something else or change their face or make them say things I've never been said before and that also was was on there however the internet I took this kind of system and next teller we did the pace to an incredible speed where those things actually goes like wildfire and I'm not talking about California I'm talking about breaking the forest tiger okay so but we see are talking about the fact that's a simple thing and I have a couple of examples which you say oh my god really so a couple of things that could happen and it actually spread like crazy okay

so let's just look at Facebook you know there's some impacts and some dire consequences in their Facebook lost a hundred twenty million dollars in a single day because people were saying hey there's too many conspiracies going on on Facebook people saw that stock dropped okay so and when we're trying to make a system to detect what a fake news is engineers cannot create rules or machine learning in such a way that humans cannot open Karina mon-sol say what is good what is bad not sure example of techniques that have been used to try to define those things and what are the criteria is behind that that allows us to say this is goodness in fact this is real to

think and the fact that we have such a vast amount of information available to us what's going to come out if you ask Google Imogen I'd like to find out if in that case the example here okay so I'd like to find out if I use in Africa see to cure cancer is there something out there that we could use the Africans seem to have cancer and I bet that if you do a search on Google and you punch that in here we have some result that say absolutely everybody can actually can cure cancer okay it doesn't mean it it's true it's just it's there how do you need at this point to use some critical thinking okay a

potentially through some fact-checking and also we have tools after a disposition to actually find out what's good and what's not it's a real problem I don't know people you know if any of you ever go to comment pizzeria story in the States they come in Caesarea you've indicated so the commitments area story was that there was a ray of pedophiles that was going around it were basically going in using the Osorio to do their deed and of course it was faking it was it was not something truth was a conspiracy theory it existed somewhere at some point the guy who walked in that to zero in any ar-15 wanting to kill those pedophiles which obviously did not

exist and we're not there but this is at the point where at people look at those stories believe them and act upon it and the availability of weapons the ability of the information makes it a very dangerous thing it's a real problem if you look here this is the the actual chart from Facebook during the election time in prediction time you have something they call notation time but all the time please look it's called something called engagement engagement this number of time you repose an article a number of times you to a comment or light or something that's number of engagement this is pretty telling during that August election date fake news at eight point seven million

engagement compared to real news at seven point three million engagement there was more engagement more people commenting sharing distributing fake news that the real thing and here some of the examples it's probably difficult for you in the back to read but things like it's over Allari Isis email just leaked and it's worse than everyone could imagine big news some people distributed them to get information Pope Francis shocks world endorses Donald Trump for president readings a statement again lots of post Reno doing of this pretty scary then you do have the very obvious one anybody read at some point just because it was so funny there Weekly World News where aliens actually come in and basically are

providing is the abduction that the Elvis is still alive and the latest one here is that UFO sparked for a spot perfect I like this one here I'll just sell your soul to the devil okay yeah step-by-step instruction and the most important one is the last one remember to defend the life extension clause because hey you know if you sell your soul you want that guarantee of 300 years of youthful life instead of just seven eighty eight right so this is easy to spot this is crack this is this is a joke this is kind of satire you did it's written dumb so it's true right it's the exact some others not so much there is a science on the internet

called effectiveness generator so you can go there and create your own figures so you specify what kind of background you want you can specify the headline and it will provide you a bunch of headlines it would even ferment you the source like in this case CBS - news s us the lieutenant comes from his trusted source succuss B comes from a trusted source and in this case the news is this one striving cheeto driven by extension by unregulated snack food industry ok this one is kind of stupid ok so you could be able to spud that fairly easily some of their from the same site not so much okay like this one here and it's

not I expected to get 16 inches of snowfall within the next week I wasn't September probably not happening or this guy traces of stds found in this western Swisher sweet sugar factory really or a guy that's dead that now is found a lion with little peep and magnet or in Cuba so you can't create those type of things and you can create the new if you don't if you're not too far-fetched you can make it believable then there are possibility to make that available so when people believe is that when paid no sorry make real news in life this is a guy that decided to create a smaller thing and he heard at some point some

the phrase bucket the losers he said oh my god braylee it was something that he said that you said that there was supposed a secret transcript of a Hillary Clinton speech given to Goldman Sachs board which claimed that Clinton had caused Bernie Sanders supporters a bucket of losers so he went on the news he went on Fox News okay and it would have been the scoop of a lifetime the problem is that that guy made the whole thing up people believe that Fox News picked it up maybe not the most reliable source in the world but still Fox News picked it up and they had to go and actually issued an on-air apology for reporting it it's pretty bad this

one this actually goes to the top so instead of state we know we said that internet will make those things happen much much faster so in this case steep speed takes precedence over truth so there's this guy copy recover the guy has 40 Twitter's forward that 4 million 40 Twitter's hope he actually decided to post it tyrion twittered which was shared 16,000 times on Twitter's and 350,000 times on Facebook the 30 was the falling the day after the election there was an anti-trump protest in Austin Texas Wow and he saw buses not far from where the protest was actually happening and he said my god this is where the protests are coming in it's a fake

protest it's not real it's not as organic as these thing those are all the buses and the people came in from Saddam okay so reddit community from mr. Trump reposted breaking they found the buses and that's the thing they found the buses okay it would just line up blocks away from the Austin process protest then it prompted over 300 comments those three comers one of some of them believed the protest on George Soros who is George Soros is the philanthropist billionaire that it is a kind of a left-wing person and now they were saying that he paid to get those people during the protest yeah and of course the president cleared like he does every

day just had a very open a successful percentage selection now professional protesters and cited by the media are protesting very unfair it went to the president that guy about 40 Twitter's formers that's all you had right so it doesn't stay there so a fox reporter called the bus convicted and asked bluntly where you involve in those test you know we'd like to know if you were involved in the actual purpose and they said no we were not so far step not okay me we're not that did not stop there rumor it continued to fly so other Megazine therefore your public they found the busted as dozens lined up blocks away from Milton if you protester

in us engineer know that we're gonna in blah blah blah blah blah they continued Tucker which was the original person they lead to this post not it up okay so the gateway pundits and other right-wing magazine said Trump figures and picked up protester were bused to us them this includes also a mentioned of lowest money now we're talking about specifically the money of Sorrows which of course never happened there either okay so you have the Gateway funded article right here and finally is a company called pebble software that little software had there and he won't kick off whether we're bringing six fifteen or sixteen fifteen thousand people into a conference and then we're using those buses to carry people back

and forth to the Contras so capital suffers a hey those are our people that are people had nothing to do with the protest you were just going to the conference center that was next to it okay so everybody thought that and that's really clear what they said so that was their statement in at the end mr. Tucker which is just his very small following Venice facts wrong he goes facts wrong he presumes something it did not fact check he posted it it went viral so from forty Twitter people it went to the president was posted three hundred fifty thousand times onto Facebook with this kind of result and I bet that today this rumor is still going

around even it has been debunked even if it's been proven not true it's still probably still going around pretty bad so that's what about this one this is actually cool one we go to a different level of

[Music]

[Music]

[Music]

[Music]

[Music]

[Music]

[Music]

pretty cool video and the sad part about this is that of course it's a fake video that the University migaila produced on purpose it did that on purpose because you know dimension Jerry here is is that the McGill University office for science and society and he was tired to see cure cancer cure from whatever happy Jews that exists somewhere and this and that thing I think which are all fake and taken from very dubious kind of material source and you said Wow you know let's create one know this job is separating sense from nonsense that's what he does so he published a video hoping to get about 10,000 G's this is what is goal because it's actually fairly close

environment that make a university well in July you add over 10 million views on this video but that's not the good part the good part is people still delete the first part of the video and it's soon promoted as a cure for cancer that does not exist that he made up so people will believe this you know everything you talk about the power of persuasion the nice music going on it's upbeat it's cool you know it has all the right words it's banned by the FDA something doctors allows you to talk about it's a natural thing all of this sounds really really good when you put it together it's fake like this okay so be aware of the fact

it would be emotionally a manipulator that's the whole point which faith is when you get to pick that was when you look at the regular news Rihanna knows our actual they tell the things the way they are when you get to fake news it's very emotional they try to play on your emotions for you to have clickbait kind of thing say click here to go to the next step that's what they do so you need to do fact-checking but also have critical thinking okay so when you see something don't think in that space value as is there's a couple of things of course consider the source when did coming from that from the onion of the New York Times okay is if they

don't read beyond so advise may be outrageous but what do they really mean is there any other sources and the other thing that are related to this check the author easier no doubt there are not that be real okay supporting sources are there you know some other links and one of the thing has been found is that if the link is that long chances are not good okay typically the destroyed link the more to the point you probably have a better chance of thing okay is it a joke could be check the date actually on my facebook yesterday one of my aunt published some things that oh my god don't buy the following products they're

all owned by Monsanto it's terrible don't go there it's awful then you start looking of course you go to different places to try to find out whatever the source of those things are and you find out the first the article was published in 2003 okay some of the companies were misspelled they were not actually right and none of them are really owned by Monsanto they may use someone said to a products but the gist of it is that not exactly true it doesn't exactly recent okay so check the best check your biases that's actually important too because we want to believe stuff we want to believe it and we'll basically kind of do everything we can to say attach to

because I built in it it doesn't mean that it's true because you believe in it okay so ask the expert go and contact somebody cut off consult some kind of site that will actually give you ideas it was a test that was done and he asked a bunch of people there's about 100 and somewhat people I think in that test and what they did is the I usually ask we'd like you to go to minimum wage calm can you evaluate that side that's all the exit go and tell me what you think of the site minimum wage calm it's real site by the way so it's not something and then put there for testing or you know it's a

real site made above Wacom so you go there and you have things immuno agent this is fact you know and you have blogs and have loves about different things so about the peoples of the earth fact checkers there was some University students and Stanford it was also experience so there were kind of a three different kind of classes of people looking at this well the fact checker all guided hundred percent right that's wrong there's something wrong with this story and the reason why it's wrong because that website or that site is backed up by a PR firm that represent the restaurant industry which is well known to net want to increase the minimum wage so all the information into

minimum wage comm goes against the phagocytic Eve all bunch of reasons not to say go on that's a good it's not a good angle destroyed economy it's not good for people you know you don't really gain anything again blah blah blah okay but if you just look at this well the permission seems to be okay however for people like sixty percent of the historian found out that this was bad and only forty percent of the students thought this was bad okay so acting is very important critical thinking of this is very important stops I'm sure most of you must have gone on the snow sand somewhere there I hope ok smelted a great source you know

companies to go there was a new move came out the great a new [ __ ] mm-hmm think you listen over and over at Christmastime right so they were actually an ad in Walmart that showed the Grinch with the pistachio app the pistachio read this great new salty just like my nuts the statue acts you say oh my god this is the kind of its children thing this is bad well obviously this was modified and faked the real ad is this one green install TV just like me that is a little bit more acceptable and that's the end of the bridge okay but what are the tell-tale sign on this first one well first you can't see ahead

of his twitter feed at the top the funds don't necessarily align know what to tell kind of time to begin with but you know lots of people believe that's in my book if you believe no we have to ask in a Walmart to remove those ads because it's actually terrible only a couple of weeks ago this is recent okay so go to Snopes and guess web so either snow it's going to get everything right they can get everything because yeah they're do have to do it with humans so you've been to have to look at this and it takes a while for those people to detect those things so you can have an army of people

looking at this but it may not be able to detect everything okay so sometimes it takes a bit of time they say thanks gonna say a day two day three days before they actually find something well guess what it can go viral in much less time than that but still at the source to actually go unchecked so MIT try to create a system to detect figures so they're still are working on it and right now they take it takes the system the electives have about 150 articles to find out if this is true or not if this is a good article or not if you have a 60 to 70 percent catch rate pretty well okay because as many things

in there that are very subjective and this is where it becomes a little bit difficult so one of the thing to do look is for example this there cycler be patient if the site has been known to publish fake news there's a chance it will do it again okay even if you have good stuff in it if you already published fake news then it creates a bad score for that site you know going this the other thing is that typically fake news they can sense before our subject and that's objective within our hyperbola is subjective and emotional so they will try to entice you to be mad to click on something to do something about it it's

not just straight news so that's not really a big indicator that something is wrong and the term of bias so they have to look at bias and nothing we kind of kind of content the type of news those have to look at the bikes so if somebody is biased to the left then the content will be different so the left-leaning concept will be kind of using harm and care fairness and reciprocity so that's kind of what you see can become their last news mania on the right side you'll see more like loyalty Authority thank God and those aren't biases so depending on your bias it doesn't mean that it's fake or not it just means that if you

see those kind of biases you know this is left-leaning if you see the other biases you're probably pretty sure it's right to me so it gives you an idea of where the biases from that bias then you can actually continue and look at from a structure okay actually the the actual content of the news itself and if you have like I said complex directory structure when you have a link or something chances are not so good okay there's typically associated with less reliable source but MIT has a open source data set with over a thousand news source so those are excites has a new site that they actually using to add their machine learning understand but

right now 60 to 70 percent efficient so they still in development to find out what's going on there of course they took the example your Infowars right so the truth above the crisis sector holds a parking hi then you see you know those are worried associated false flag demand comes courtesy those are the type of things pretty good chance that's pretty bad you don't need to go very far to understand that this is no good okay that is also this is a free app by the way you can download this you need a docker container you can load it in a docker container and run the different URL and news that you have to see what

they think this is actually pretty good it's a guy that decided to create a fake news program and he started to look at the different ways to look at fakers so you try to categorize them there's a lot of kind of news so you have real faith satire misleading opinions retweeting together so we started with a sentiment so basically how does it feel how does other words relate to each other and how what kind of cinnamon comes out of it so he started to death and it failed miserably with a wet paper it didn't work if you just look at a sentiment well know some good new was also extended so this is not visalia quite

particular okay so you continually tried to use natural language processing to try to categorize the data somehow is okay this is good this is a job with satire this is real this is this and again that actually failed miserably and I wasn't very good so he said well you know what why don't we look it'll be around we know what the good news are so let's look at good news and see what the difference between the good news and anything else that we're looking at so if we know the good news coming from reliable source aren't good if we compare it to the structure the way they actually have the sentiment the way they're categorized then we may be able

to come to some kind of consensus as to is it good or bad and it did that and when he did this he came up with a result at about 95 percent accuracy that's too if it's fake things or not it's not perfect it's actually pretty that good so this is free it says something you can you can actually download and install okay so he found that if an article is special in to the point that is a good article if it's long he was all bunch of different words and it will basically kind of deviate a lot from a regular that's not this this is start to be in the realm of nature okay so he created

that app called fake bucks it's free for personal use like I said I use a darker image and you can just drop the content in the actual document will come out of that say eight it's probably there that however technology may not be able to fix this right now unfortunately there's a lot of human involvement in it that's why when I started they said you have you know a good part to play in their technology can spot a few things it will never be perfect and even if a technology can detect some of that stuff guess what that's somebody that will find a way to go over and above that that makes you should get nobody can

that detect it then will be in the same conundrum that you were in the beginning okay so when you look at those things you know humans can't even agree on sometimes in the results is this really true is it pretty not true your bias comes into play you wanted to leave this therefore it's good but the other guy wants to believe this the points back and you can't get the consensus abrini is good or bad even if you have the proper facts sometimes okay so coming sense still prevents if it sounds too good to be true probably is okay I haven't seen yet the actual Saudi prince came to me and says I'm going to

give you fifty eight million dollars you only have to be too grand to process it didn't happen it okay so we have to be critical we have to understand what's out there make their research if something sounds fishy do some research okay who are yearning or what is the source waves coming from what is it they're going to tell me was the authors the author of somebody common that we know about and yes it takes much longer unfortunately that's the way it has to be right now there's too many things that are being propagated distributed retweeted reposted that don't make any sense and create stories like the common concern or like the fake buses that actually you know

that buses for the protesters all those things are acting on a daily basis so be after kind of critical thinking and do some fact-checking when you do stuff that's it for me thank you any questions on this I was very subjective but yes well we're working on a bunch of different things whether or not things will be coming as a product unknown again right now is it's extremely volatile it's extremely subjective so it's it's not easy to create something that will be reliable but yes we do have people working on a bunch of different technologies left and right this thank you

Thank You Savannah pepper name devastation after three o'clock