“Enragement is engagement,” and the algorithms are helping disinformation speed around the globe these days based on what enrages/engages users the most.
I’d say the point about the algorithms is indirectly true. And now we’ll go down that rabbit hole.
Facebook (and most other advertisement-funded, social media platforms) were built on algorithms designed to drive user engagement for collecting user information to sell their targeted advertisement capabilities to businesses. That drove their algorithms to serve advertisements to users based on user demographics and engagement with similar adds across multiple websites. The social media content they served was based on your connections to others and content they created and shared.
TikTok has revolutionized that in their approach. They serve content in a similar way as Facebook serves advertisements, but they take it much further. TikTok generates a large amount of metadata about content that’s uploaded to its site. Leveraging computer vision (CV) and large language model (LLM) tools, they extrapolate the metadata directly from the content. The algorithms detect the subject, context, and sentiment of the content, as well as things like the location, demographics, etc of the personalities involved. And then they map each piece of content in a multidimensional similarity graph according to each of those metadata attributes. Additionally, they record user interactions with content they served creating a similar multi-dimensional array. And the serve content through a 2-dimensional nearest-neighbor algorithm that combines these 2 multi-dimensional arrays.
“Engagement is engagement” drives many users to engage more with enraging content, leading to an algorithmic bias that leads to digital echo chambers furthering the effect of the previously discussed disinformation platforms.
This is what makes TikTok the most effective disinformation ecosystem on its own, and even worse when the CCP decides to weaponize it. Some folks would argue they have at least tested in this current Israel-Palestine conflict.
Thank you. I’ve got a quality mentor in Will who taught me half of what I know about writing; mostly about the practical use of commas and periods for readability. 😁
Will and James, I am saving this for rereading. The comments alone are so informative and thought-provoking. As a boomer-layperson, I sit in the midst of so much propaganda; Many of us need suggestions/guidance to land somewhere close to the truth. I am exhausted with the "headline throwers". Help.
James, thanks much for this analysis. I really appreciate the analogy to your experience in Afghanistan. You write very well and provide very cogent analysis.
I think you are right on target re: disinformation, but the conundrum is how do we, here in the US, and the IDF in Gaza, deal with this reality? No doubt, right now, Hamas is winning the "propaganda war." Israel is rapidly losing its goodwill in the internation community and is clearly hated in Gaza.
If you cannot beat them, join them. Not in disinformation, but in technique. Disinformation goes viral because humans share it more often, per the MIT study. The IDF, or other Israeli agencies, COULD overwhelm the “viral” spread of disinformation by producing their own massive streams of “anti-bodies” or information operation based in a more truthful narrative. So far, they’ve expected to compete on level footing with disinformation by producing competing content and setting it free in the wild. They COULD use bots to drive up exposure to their counter content. But how do you achieve that? You reverse engineer the Russian disinformation machine and remake something in its image but to push fact based content instead of lies.
The best example of what the Russians did in their 2016 full-throated influence campaign to interfere in the US democratic process was the @TN_GOP Twitter account. The false account featured anti-Muslims, anti-immigrant, and other racist content that had obvious grammatical and English errors but it collected a following over 10x the actual TN GOP party’s legitimate page. And what’s more, the Trump campaign officials up to and including Donald Trump Jr. even reshared many of their tweets. The state party submitted multiple requests to have the account deleted, but it wasn’t until 2017 that Twitter took down the account. By that time, it had outlasted its true usefulness to the Russian disinformation machine as they pivoted to upcoming elections in Europe.
Israel COULD find the approach of building personas as a form of Information Warfare counter-messaging “platforms” that infiltrate the digital information sphere where those they are most likely to influence with their truth-based content. This is nearly impossible to do in the moment of need as it takes months if not years to build the reach needed to leverage the personas as information operation platforms, but it can be done with good planning.
This Information Warfare approach is akin in my mind to its Electronic Warfare cousin, which to be effective must have 3 components: frequency, access, and power.
For Information Warfare this looks likes frequency in terms of channel or more effectively the right groups across various Social Media platforms around the globe. TikTok, Twitter, Facebook, and YouTube would be the biggest 4 for many Western markets, but there’s probably unique channels for Africa, Russia, the Middle East, and Asia with some unique to individual countries.
Access means the personas need to be widespread within those platforms. Much like the @TENN_GOP Twitter account outpaced the legitimate in followers, Information Warfare platforms need to have access to a large portion of their intended audience to effectively disseminate their content. This is the part that takes the most time. Those personas have to be built over months or years without being detected as fake, and they have to both “hide in plain sight” as well as grow their “access” deliberately.
Power is volume of disparate copies of the same narrative. The message campaigns when initiated have to rise above the noise level, and “dominate the sphere” where they exist to over power the message of the disinformation machines attempting to do the same. This is where I believe the Information Warfare platforms have to begin to interact. If humans aren’t going to spread the truth, the platforms must be capable of doing so themselves. And in the age of Generative AI this is getting easier each day.
The implementation of this would be nuanced to be both achievable and ethical. And, it must be cultivated deliberately over time. But it might just be the best chance truth has to compete with disinformation on a global scale. Because humankind is proving that we seem to be incapable of the necessary critical thought to overcome disinformation unassisted. That’s just my opinion.
The criticism I've seen of the IDF that is compelling comes not from posts about Isreali atrocities, but by military bloggers talking about how the IDF has changed recently in its tactics and strategies. The Substack Fire for Effect had a post on the IDF Decisive Victory strategy, and how it was conflating tactics with strategy. One main point was that it was taking the human element out of decision making and focusing on target servicing. Then add to that the recent story of how the IDF uses AI (Lavender) to generate large target lists that are not scrutinized, just serviced. Again, commanders in the IDF have taken human decision making out of the loop.
I wonder which IDF you think is operating here:
The IDF that behaves much like you did in Afghanistan where targets were evaluated by a group of command staff before they were struck.
...Or...
The IDF who has changed its strategy to destroying targets as fast as possible, with target lists being created by AI that no one bothers to scrutinize anymore because that isn't part of the new tactics or strategy in place.
I've been very moved by what you did in Afghanistan, and how you executed your mission. I also wonder if you are putting yourself in IDF shoes, thinking they are fighting like you did, when those shoes don't fit.
I know that the IDF used to consider civilian casualties... but if they are specifically targeting entire families of anyone who may be loosely associated with Hamas now, that consideration is gone. Not to mention all the other families who live in the high density housing around that family. Now the stories of IDF atrocities that wouldn't have made sense before, have some kind of rational basis. The source of it isn't "Those dastardly Jews want to genocide", the source is "There is a very effective military that has removed its humanity on purpose, and as a consequence is letting itself behave in a way that would be unimaginable if people actually made these decisions."
You say repeatedly the IDF is going out of its way to not harm civilians... like they used to do... Is that what they are still saying, or is that what they are still doing?
I'm unsure who the comment is directed towards: James or me. But I'll field it. I think using AI to conduct targeting campaigns is the way of the future. The USAF is moving in that direction too. I would also add that not all targets were evaluated by a group of command staff before they were struck. Sometimes they were, especially if it was a mission intended to strike HVTs. However, armed drones also conducted ISR for TICs. And would fire and forget. For all of the talk about Israelis not considering civilian casualties (which they do. I encourage you to read John Spencer on this topic, who wrote the book on the subject.) I hate using body counts as a metric for reasons I've laid out, but if you take Hamas' tally of about 35,000 and the IDF kill count of ~13K, then you're talking about a 1:3 kill ratio of combatants vs. noncombatants. That is unheard of in urban warfare. I would also add that while we talk about how accurate the IDF needs to be, Hamas continues to rape abducted Israeli women. That is happening right now - as we speak. They are also firing IDF indiscriminately, hoping to strike civilians. The IDF has done quite a bit to lessen civilian casualties. Those numbers are dropping because they're learning as they fight. We shall see if it holds up or not. War is fluid. They might very well lose.
I would add that I’ve read some assertions that the IDF is just servicing AI-nominated targets, but have yet to be presented any evidence that supports it. The IDF fired 2 commanders as a result of the investigation into the death of these aid workers. You don’t fire humans when you’re entirely reliant on the AI. You fire humans when the humans made bad decisions.
I’ll buy that the IDF may be pushing the envelope with AI-driven targeting and that they may not be properly accounting for bias in their algorithms. That’s a legitimate ethical concern when it comes to military use of AI. I would also argue that the culture can be just as bias due to those charged with targeting the adversary being under pressure (command induced or self induced) to make progress towards defeating Hamas. I watched that many times in my experience where folks wanted to kill, and that desire led to some bad assessments and misidentifications; especially when those performing the analysis had never been outside the wire in Afghanistan, and had only ever seen it from a screen. I always found it harder to want to kill the person in the pixels having broken bread with Afghans and knowing them as something other than a foreign people in some far off land.
And one additional point after reading more bloviating personalities (present company not included) on this topic: Hamas members families shouldn’t be regarded as “innocent civilians” even if they aren’t combatants by the Law of War. Those personalities seeking to undermine the facts and assert that “Hamas isn’t using human shields” because the IDF is supposedly conducting strikes when the Hamas members are return to their homes in the evening where their families are “innocent civilians” killed in the strikes. ISIS had no problem targeting the families of US military members, or at the very least calling for them to be targeted by their sympathizers in lone wolf attacks aimed at subverting the will of the US military members to continue drone strikes decimating their ranks in Syria & Iraq. The US had to continuously monitor the ISIS kill lists where they “doxed” US service members and their families they “cleared for targeting.” I recall hearing of at least one military member being reassigned to a different location after his name and address were posted. Luckily, many of the names and/or addresses were incorrect or at the very least no longer valid. So, the families of Hamas members aren’t combatants, but they might not have the same consideration as innocent civilians when Hamas chooses to fight their war for the total annihilation of Israel from their own homes.
You mentioned the “PR problem” that the IDF is facing. I would argue a large part of this issue is because it’s being reported and discussed in a false context, an alternative universe, where it’s often compared to the numbers, effects, and paced of the US’s protracted wars in Iraq and Afghanistan. The contrast is that the US paced our wars against our adversaries half a globe away with limited reach to attack the homeland. Israel is fighting their war to destroy an adversary “inside their doorway” and they don’t have the luxury of time at their disposal.
We might not get our news from social media (alone at the very least) but the news we get is arguably being drawn to where their audience is, and that’s growing increasingly more bias in the age of social media-based disinformation. A great example of this is Fox News knowing spreading false information about the 2020 election because the former president’s use of Twitter as his bully pulpit to spread his big lie. As his base followed him further into the alternate universe he erected through his disinformation, FoxNews chased the crowd completely into the Dominion lawsuit.
Much of the “factual reporting” I’ve read by Israel’s critics are a stream of statistics and facts encased is a soft narrative and served to an audience with designed inherent bias because of this social media disinformation.
And you are making a difference, both of you. Keep writing.
“Enragement is engagement,” and the algorithms are helping disinformation speed around the globe these days based on what enrages/engages users the most.
I’d say the point about the algorithms is indirectly true. And now we’ll go down that rabbit hole.
Facebook (and most other advertisement-funded, social media platforms) were built on algorithms designed to drive user engagement for collecting user information to sell their targeted advertisement capabilities to businesses. That drove their algorithms to serve advertisements to users based on user demographics and engagement with similar adds across multiple websites. The social media content they served was based on your connections to others and content they created and shared.
TikTok has revolutionized that in their approach. They serve content in a similar way as Facebook serves advertisements, but they take it much further. TikTok generates a large amount of metadata about content that’s uploaded to its site. Leveraging computer vision (CV) and large language model (LLM) tools, they extrapolate the metadata directly from the content. The algorithms detect the subject, context, and sentiment of the content, as well as things like the location, demographics, etc of the personalities involved. And then they map each piece of content in a multidimensional similarity graph according to each of those metadata attributes. Additionally, they record user interactions with content they served creating a similar multi-dimensional array. And the serve content through a 2-dimensional nearest-neighbor algorithm that combines these 2 multi-dimensional arrays.
“Engagement is engagement” drives many users to engage more with enraging content, leading to an algorithmic bias that leads to digital echo chambers furthering the effect of the previously discussed disinformation platforms.
This is what makes TikTok the most effective disinformation ecosystem on its own, and even worse when the CCP decides to weaponize it. Some folks would argue they have at least tested in this current Israel-Palestine conflict.
You conveyed your message clearly, James.
Thank you. I’ve got a quality mentor in Will who taught me half of what I know about writing; mostly about the practical use of commas and periods for readability. 😁
Will and James, I am saving this for rereading. The comments alone are so informative and thought-provoking. As a boomer-layperson, I sit in the midst of so much propaganda; Many of us need suggestions/guidance to land somewhere close to the truth. I am exhausted with the "headline throwers". Help.
James, thanks much for this analysis. I really appreciate the analogy to your experience in Afghanistan. You write very well and provide very cogent analysis.
I think you are right on target re: disinformation, but the conundrum is how do we, here in the US, and the IDF in Gaza, deal with this reality? No doubt, right now, Hamas is winning the "propaganda war." Israel is rapidly losing its goodwill in the internation community and is clearly hated in Gaza.
If you cannot beat them, join them. Not in disinformation, but in technique. Disinformation goes viral because humans share it more often, per the MIT study. The IDF, or other Israeli agencies, COULD overwhelm the “viral” spread of disinformation by producing their own massive streams of “anti-bodies” or information operation based in a more truthful narrative. So far, they’ve expected to compete on level footing with disinformation by producing competing content and setting it free in the wild. They COULD use bots to drive up exposure to their counter content. But how do you achieve that? You reverse engineer the Russian disinformation machine and remake something in its image but to push fact based content instead of lies.
The best example of what the Russians did in their 2016 full-throated influence campaign to interfere in the US democratic process was the @TN_GOP Twitter account. The false account featured anti-Muslims, anti-immigrant, and other racist content that had obvious grammatical and English errors but it collected a following over 10x the actual TN GOP party’s legitimate page. And what’s more, the Trump campaign officials up to and including Donald Trump Jr. even reshared many of their tweets. The state party submitted multiple requests to have the account deleted, but it wasn’t until 2017 that Twitter took down the account. By that time, it had outlasted its true usefulness to the Russian disinformation machine as they pivoted to upcoming elections in Europe.
Israel COULD find the approach of building personas as a form of Information Warfare counter-messaging “platforms” that infiltrate the digital information sphere where those they are most likely to influence with their truth-based content. This is nearly impossible to do in the moment of need as it takes months if not years to build the reach needed to leverage the personas as information operation platforms, but it can be done with good planning.
This Information Warfare approach is akin in my mind to its Electronic Warfare cousin, which to be effective must have 3 components: frequency, access, and power.
For Information Warfare this looks likes frequency in terms of channel or more effectively the right groups across various Social Media platforms around the globe. TikTok, Twitter, Facebook, and YouTube would be the biggest 4 for many Western markets, but there’s probably unique channels for Africa, Russia, the Middle East, and Asia with some unique to individual countries.
Access means the personas need to be widespread within those platforms. Much like the @TENN_GOP Twitter account outpaced the legitimate in followers, Information Warfare platforms need to have access to a large portion of their intended audience to effectively disseminate their content. This is the part that takes the most time. Those personas have to be built over months or years without being detected as fake, and they have to both “hide in plain sight” as well as grow their “access” deliberately.
Power is volume of disparate copies of the same narrative. The message campaigns when initiated have to rise above the noise level, and “dominate the sphere” where they exist to over power the message of the disinformation machines attempting to do the same. This is where I believe the Information Warfare platforms have to begin to interact. If humans aren’t going to spread the truth, the platforms must be capable of doing so themselves. And in the age of Generative AI this is getting easier each day.
The implementation of this would be nuanced to be both achievable and ethical. And, it must be cultivated deliberately over time. But it might just be the best chance truth has to compete with disinformation on a global scale. Because humankind is proving that we seem to be incapable of the necessary critical thought to overcome disinformation unassisted. That’s just my opinion.
The criticism I've seen of the IDF that is compelling comes not from posts about Isreali atrocities, but by military bloggers talking about how the IDF has changed recently in its tactics and strategies. The Substack Fire for Effect had a post on the IDF Decisive Victory strategy, and how it was conflating tactics with strategy. One main point was that it was taking the human element out of decision making and focusing on target servicing. Then add to that the recent story of how the IDF uses AI (Lavender) to generate large target lists that are not scrutinized, just serviced. Again, commanders in the IDF have taken human decision making out of the loop.
I wonder which IDF you think is operating here:
The IDF that behaves much like you did in Afghanistan where targets were evaluated by a group of command staff before they were struck.
...Or...
The IDF who has changed its strategy to destroying targets as fast as possible, with target lists being created by AI that no one bothers to scrutinize anymore because that isn't part of the new tactics or strategy in place.
I've been very moved by what you did in Afghanistan, and how you executed your mission. I also wonder if you are putting yourself in IDF shoes, thinking they are fighting like you did, when those shoes don't fit.
I know that the IDF used to consider civilian casualties... but if they are specifically targeting entire families of anyone who may be loosely associated with Hamas now, that consideration is gone. Not to mention all the other families who live in the high density housing around that family. Now the stories of IDF atrocities that wouldn't have made sense before, have some kind of rational basis. The source of it isn't "Those dastardly Jews want to genocide", the source is "There is a very effective military that has removed its humanity on purpose, and as a consequence is letting itself behave in a way that would be unimaginable if people actually made these decisions."
You say repeatedly the IDF is going out of its way to not harm civilians... like they used to do... Is that what they are still saying, or is that what they are still doing?
I'm unsure who the comment is directed towards: James or me. But I'll field it. I think using AI to conduct targeting campaigns is the way of the future. The USAF is moving in that direction too. I would also add that not all targets were evaluated by a group of command staff before they were struck. Sometimes they were, especially if it was a mission intended to strike HVTs. However, armed drones also conducted ISR for TICs. And would fire and forget. For all of the talk about Israelis not considering civilian casualties (which they do. I encourage you to read John Spencer on this topic, who wrote the book on the subject.) I hate using body counts as a metric for reasons I've laid out, but if you take Hamas' tally of about 35,000 and the IDF kill count of ~13K, then you're talking about a 1:3 kill ratio of combatants vs. noncombatants. That is unheard of in urban warfare. I would also add that while we talk about how accurate the IDF needs to be, Hamas continues to rape abducted Israeli women. That is happening right now - as we speak. They are also firing IDF indiscriminately, hoping to strike civilians. The IDF has done quite a bit to lessen civilian casualties. Those numbers are dropping because they're learning as they fight. We shall see if it holds up or not. War is fluid. They might very well lose.
I would add that I’ve read some assertions that the IDF is just servicing AI-nominated targets, but have yet to be presented any evidence that supports it. The IDF fired 2 commanders as a result of the investigation into the death of these aid workers. You don’t fire humans when you’re entirely reliant on the AI. You fire humans when the humans made bad decisions.
I’ll buy that the IDF may be pushing the envelope with AI-driven targeting and that they may not be properly accounting for bias in their algorithms. That’s a legitimate ethical concern when it comes to military use of AI. I would also argue that the culture can be just as bias due to those charged with targeting the adversary being under pressure (command induced or self induced) to make progress towards defeating Hamas. I watched that many times in my experience where folks wanted to kill, and that desire led to some bad assessments and misidentifications; especially when those performing the analysis had never been outside the wire in Afghanistan, and had only ever seen it from a screen. I always found it harder to want to kill the person in the pixels having broken bread with Afghans and knowing them as something other than a foreign people in some far off land.
And one additional point after reading more bloviating personalities (present company not included) on this topic: Hamas members families shouldn’t be regarded as “innocent civilians” even if they aren’t combatants by the Law of War. Those personalities seeking to undermine the facts and assert that “Hamas isn’t using human shields” because the IDF is supposedly conducting strikes when the Hamas members are return to their homes in the evening where their families are “innocent civilians” killed in the strikes. ISIS had no problem targeting the families of US military members, or at the very least calling for them to be targeted by their sympathizers in lone wolf attacks aimed at subverting the will of the US military members to continue drone strikes decimating their ranks in Syria & Iraq. The US had to continuously monitor the ISIS kill lists where they “doxed” US service members and their families they “cleared for targeting.” I recall hearing of at least one military member being reassigned to a different location after his name and address were posted. Luckily, many of the names and/or addresses were incorrect or at the very least no longer valid. So, the families of Hamas members aren’t combatants, but they might not have the same consideration as innocent civilians when Hamas chooses to fight their war for the total annihilation of Israel from their own homes.
'Amen. War is all fun and games from an analytical perspective until you remember how it felt to get targeted by these mofos.
You mentioned the “PR problem” that the IDF is facing. I would argue a large part of this issue is because it’s being reported and discussed in a false context, an alternative universe, where it’s often compared to the numbers, effects, and paced of the US’s protracted wars in Iraq and Afghanistan. The contrast is that the US paced our wars against our adversaries half a globe away with limited reach to attack the homeland. Israel is fighting their war to destroy an adversary “inside their doorway” and they don’t have the luxury of time at their disposal.
We might not get our news from social media (alone at the very least) but the news we get is arguably being drawn to where their audience is, and that’s growing increasingly more bias in the age of social media-based disinformation. A great example of this is Fox News knowing spreading false information about the 2020 election because the former president’s use of Twitter as his bully pulpit to spread his big lie. As his base followed him further into the alternate universe he erected through his disinformation, FoxNews chased the crowd completely into the Dominion lawsuit.
Much of the “factual reporting” I’ve read by Israel’s critics are a stream of statistics and facts encased is a soft narrative and served to an audience with designed inherent bias because of this social media disinformation.