Deep fakes in Warfare yep according to a Report by The Brookings institution we Could one day see it happening the Authors of the report claim that deep Fakes will be used by Foreign Adversaries for disinformation and Misinformation campaigns and they Propose various policies to address this Threat so today I'm going to summarize This report tell you what it says in Simple terms and look at why this new Technology could be used to justify Unprecedented internet censorship and Control Good The report I'll be summarizing today is Titled quote deep fakes and International conflict I'll leave a link To it in the description if you're Interested in reading the whole thing And I'll reiterate that the Brookings Institution is very influential so What's in this report should not be Taken lightly Now the report begins with an Introduction which talks about the use Of deep fakes following Russia's Invasion of Ukraine just a few days After the invasion someone presumably The Russian government posted a deep Fake of Ukrainian President Vladimir Zielinski telling ukrainians to Surrender although this kind of Information Warfare has historically
Come hand in hand with Kinetic Warfare The authors point out that this was the First quote high-profile use of a deep Fake in an international conflict and it Won't be the last that's because of the Rapid advancement of Technology mainly AI as such the authors call on Governments to quote develop robust Systems for monitoring and Authenticating both public and private Messages in real time not only that but The authors go as far as recommending That governments assess when it's Appropriate to use deep fakes for Offensive purposes They go on to Proclaim that from here on Out governments must assume that their Adversaries have access to deep fake Technologies that make it possible to Perfectly mimic any leader politician or Soldier the authors also say that one Method of mitigation could be so-called Pre-bunking now if you watched our video About the world economic forum's DeVos Conference you'll know that pre-bunking Means making it impossible to post Certain kinds of information Pre-bunking also includes suppressing The reach of certain kinds of Information which is effectively the Same as censorship The authors then cautioned that creating Appropriate anti-deep fake policies as Well as knowing when to use deep fakes
As an offensive weapon requires that Policy makers understand the underlying Technology to that end the author's call For the creation of a quote deep fakes Equities process They reveal that this benign sounding Measure has its roots in another Military measure called the quote Vulnerabilities equities process this Military process basically involves Deciding whether it's appropriate to Exploit a fatal technological flaw in a Hostile country's digital infrastructure Yikes the second part of the report Provides an explanation of what deep Fakes are and how they work the authors Start by explaining that deep fakes are Only possible because of recent Advancements in machine learning these Advancements are primarily due to the Availability of data and much more Powerful Hardware They explain that this has made it Possible for machine learning to Essentially evolve into something called Deep learning now you can think of Machine learning as being the ability to Imitate whereas deep learning is the Ability to replicate the authors also Explain that the deep in deep learning Is why deep fakes are called Deep fakes Now what you're looking at here is a Diagram which shows how deep fakes are Made I must admit that the author's
Explanation flew a bit over my head so Feel free to correct me in the comments Section To my understanding a computer program Called a generator creates fake images a Second computer program called a Discriminator Compares these fake images With real images The generator generates as many images As it can until the discriminator can't Distinguish between the real and fake Images this is called a generative Adversarial Network or Gan and the Authors explained that the technology Has been around since 2014. the Difference today is that Gans exist for All different kinds of data and again Have access to more data and more Powerful Hardware to make the fake Output seem even more real the authors Go on to provide an example of a gan Called tread short for terrorism Reduction with artificial intelligence Deep fakes which was developed by a us University military researchers use Tread to make a deep fake of an Islamic Terrorist saying something that he never Said not very subtle What's crazy is that the authors admit That tread was developed using open Source gam code they caution that anyone With lots of machine learning experience Lots of data and around ten thousand Dollars worth of powerful computer
Hardware can create a realistic deep Fake using tread and other Gans what's Also crazy is that the authors admit That deep fakes have evolved to a point That they're next to impossible for the Average person to spot this has resulted In the creation of deep fake detection Technologies which look for Inconsistencies in facial expressions Eye movements and other things what's Even crazier is that the authors admit That Gan developers will use new deep Fake detection Technologies to train Their discriminator programs to be even Better this results in what the authors Call a quote cat and mouse game similar To that of malware detection with Malware evolving over time Anyways the third part of the report Talks about the potential uses of deep Fakes in international conflicts the Authors start by recounting how Russia Allegedly undermined Hillary Clinton's Presidential campaign in 2016 using Disinformation and misinformation Resulting in Donald Trump's Victory the Authors reveal that disinformation and Misinformation spreads faster than true Information but I suppose that depends On your definition of these terms in any Case the authors accurately underscore That deep fakes will soon make it Impossible to know what's real and What's not
The authors then provide a few Historical examples of how false Information was used in Warfare some of These are quite amusing such as quote Octavian used poetry and coins to spread The impression that Mark Anthony was a Philandering drunk must have been some Pretty convincing poems Now some of the other examples the Authors provide are a bit more Contentious such as the KGB claiming That U.S President John F Kennedy was Assassinated by U.S intelligence I Suggest reading JFK's address to the American newspaper Publishers Association on the 27th of April 1961 Link down below Now the authors go on to detail the Different ways that deep fakes could be Used in Warfare going forward deep fakes Could be used to legitimize Wars and Revolutions they could be used to Falsify orders they could be used to Cause social confusion they could be Used to confuse the military Deep fakes could be used to undermine Popular support in a government's Military efforts they could be used to Polarize societies they could be used to Divide allies and they could be used to Discredit leaders each deep fake use Case comes with lots of examples and the Last use case is the longest this Doesn't seem to be a coincidence because
Trust in institutions is collapsing if You watched our video about online Censorship you'll know this collapse in Trust is universal more importantly the Main cause of this collapse in trust was The pandemic restrictions that were Implemented by governments now believe It or not but the authors then have the Audacity to State quote Although the focus above is on how Countries might use deep fakes against The United States and its allies deep Fakes might be used against U.S Adversaries and other threats in order To protect the international order In other words if countries that are Hostile to the United States and its Allies use deep fakes then it's bad but If the United States and its allies use Deep fakes against these hostile Countries then that's good to be clear I'm not on anyone's side here I honestly Think all sides have equal levels of Corruption regardless the authors Acknowledge that deep fakes don't always Work and even when they do they will be Fairly easy for governments to discredit They also add that as public awareness Of deep fakes increases then there will Be a sort of Baseline skepticism of the Information that's being shown online This obviously creates another problem Which is that the average person may Also become skeptical of official
Information as a result and this ties Into the fourth part of the report which Contains a long list of policy Recommendations related to deep fakes The authors cover both defensive and Offensive policies in terms of defensive Policy the authors admit that creating Deep fake detection Technologies won't Work in the long run because deep fake Developers will inevitably upgrade their Gans to avoid detection that leaves only One option and that's the authentication And verification of digital content this Begs the question of what such Authentication and verification would Entail and the answer is probably kyc if You watched our video about online Censorship you'll know that governments Worldwide are introducing laws that will Require you to complete kyc to use Social media in the name of safety the Authors also suggest making sure that Any new information that comes out isn't Immediately accepted as true until it's Been verified by multiple third parties They seem to advocate for slowing down How quickly information is shared online So that the fact checkers can do this Verification work in terms of offensive Policy meanwhile the authors advise Against using deep fakes for domestic Propaganda purposes because quote a Botched manipulation of one news story Might discredit government efforts to
Address hundreds of other pieces of news To that I say look up operation Mockingbird while you're at it check out Our video about social media censorship Link will be in the description I Digress Now the authors also advise against Using deep fakes for intelligence Operations because they carry Unspecified risks they then say this Quote Consider how the United States and other Democracies have used public health Programs as cover for their operatives That's not suspicious at all In all seriousness the authors are aware That democratic governments will Nonetheless use deep fakes for both Defensive and offensive purposes given This fact they suggest the United Nations set deep fake standards and that The U.S introduced the aforementioned Deep fake equities process now the Remainder of the report is a deep dive Into how deep fakes work including Classification algorithms and deep Learning what's funny is they use a Picture of one of the authors to Illustrate how deep fakes classify Information it's worth a read if you Have the time and the brain capacity so This brings me to the big question and That's why the emergence of deep fakes Could be used to justify unprecedented
Internet censorship and control well if The last few minutes didn't make it Clear enough deep fakes will make it Next to impossible to know what Information is true or false online This sounds silly Until you realize that AI is growing at an exponential rate Chat GPT is just the tip of the iceberg Soon there will be dozens of advanced AIS each with their own biases and Abilities it's only a matter of time Before the average person has access to This powerful Tech too If you watched our video about chat GPT You'll know that one of the most Powerful things it can do is write code I reckon it won't be long before people Start using chat GPT style AIS to create Advanced gas that can produce deep fakes That are realer than the real thing when This happens the internet will become Flooded with misinformation and Disinformation and not just Controversial facts that the people in Power don't like we could see serious Domestic and international conflicts Arise which could destroy the world as We know it this is why I'm starting to Become convinced that governments around The world will do something to slow down The development of both Ai and deep Fakes in theory they could announce an Emergency pause on their use and Development but in practice that would
Be next to impossible to enforce Alternatively governments could engage In the Cyber attack the elites keep Talking about this would take everything Offline and allow only verified users And content to reappear as exciting as Such an option would be it would clearly Come with its own implementation issues So this leaves only one other option and That's requiring All Digital content to Be authenticated and verified as per the Author's advice recall that this is code For forcing everyone to do an identity Check if they want to post something Online this would give the government Total control of information online the Worst part is that this approach Wouldn't work to stop deep fakes either That's simply because governments would Continue to post deep fakes from Authorized accounts for domestic Propaganda purposes this propaganda Would likely be related to going to war Or say fighting a pandemic to add insult To injury anyone who claims that the Official information being given by Governments online is fake would likely Be banned if not outright sent to prison And if they tried to present evidence That the government was publishing fake Information this evidence would be Branded as a deep fake So what's the solution then well easy Nfts on trustless decentralized
Permissionless and publicly viewable Cryptocurrency blockchains This solution makes the most sense Because it would make it possible to Preserve privacy online while verifying The authenticity of information online Instead of having to complete kyc to use Social media you could just use it as You do now all information you generate Could be turned into nfts which only you Control this would make it easy to Verify that any information about you Came from you and not from someone else From where I'm standing authenticating And verifying information that's posted Online will be the killer use case for Nfts the only caveat is that the Underlying blockchain must have the Qualities I just mentioned a few moments Ago otherwise you might as well just Scan your eyeball to Post online with Some luck the crypto industry will Realize that fighting deep fakes with Nfts is the way to go before governments Start exploring more dystopian methods I'm confident that this will happen and Hopefully this video will help with that Outcome so la like subscribe and share This video to do your part [Music]
