simulations – Techdirt (original) (raw)

We Ran Our Online Election Disinformation Simulation Game And There's Plenty To Be Worried About

from the people-are-sneaky dept

If you are interested in having us run Threatcast 2020, or commission some other “serious” games, for your organization or as a group event, please contact us..

Back at the end of January, you may recall that we wrote about Threatcast 2020, an in-person election disinformation brainstorming simulation that we had created last year — created in partnership between our think tank organization, the Copia Institute, and Randy Lubin of Leveraged Play. The game was developed as an in person brainstorming exercise to look at various strategies that might be used to engage in (and counter) disinformation and misinformation strategies around the 2020 election. We had hoped to run the event throughout this year.

Of course, soon after we announced it, the pandemic hit the US pretty hard, and the idea of running in-person events disappeared. The game had a variety of specific elements to it, and replacing it via Zoom just wouldn’t be the same. After it became clear that the pandemic situation would almost certainly rule out all in-person events this year, we set to making an online version of the game, which we completed a few weeks back. We’ve now run the event a few times, some for private groups, and one “showcase” event we put on just last week. The event itself was run under Chatham House rules, so we will not identify who attended or what individuals said, but I can talk a bit about what happened at the event. And, just for clarification, we had a wide range of participants — from companies, non-profits, foundations, academia, and government.

One participant who did agree to be named was famed investor Esther Dyson, who told me of the event that “It was fun and funny, but it had enough truth in it to be an amazing and eye-opening experience. This kind of simulation is exactly the preparation people need for the real world, whatever world they operate in.” She also noted her key takeaway from the event: “The most compelling message is that the chaos hackers were almost redundant in the ugly world that the two warring parties – or four warring factions – were creating for themselves and all around them. Our wish, in playing as the chaos team, was for a contested election, not a specific winner. And a final key message: it will be important to see who can bring us together – especially AFTER the election.”

The game itself involves players working in teams as various political factions — representing a broad coalition of political operatives (not as specific candidates or campaigns) — and responding to certain situational prompts (and actions by other teams) as they navigate from now through the election (and beyond). Not all of the factions are interested in supporting a happy democratic election. In the event we ran last week, there were four rounds covering the run-up to the election and the immediate aftermath of the election.

The players brought a vast array of manipulation and deception to the campaigns and created an atmosphere of paranoia, anger, and confusion. Over the course of the election, the center-right republicans turned their focus to down-ballot races, enabling the GOP to keep the Senate and retake the House of Representatives even as the Democrats won the presidency. However, Trump refused to concede defeat and the game ended with a standoff at the White House. I should note that while there is, within the game, some election modeling to see how well these strategies impact the actual election, the game is not designed to simulate (and certainly not to predict) the outcome of the election, but rather to simulate what kinds of disinformation we’ll see (across the board). Along those lines, I’ll note that the results of this simulation turned out quite different than the other Threatcast’s we have run.

Of particular interest in last week’s simulation: the amount of chaos. If 2020 has taught us anything, it’s that nothing seems off the table, and no idea is too crazy. That played out within our game as well (though, at least one of our judges noted that even some of the more “extreme” ideas presented were ones that were already playing out in real life). Another element that played out, as Esther Dyson noted above, was just how much chaos there is overall — such that some of the players (who were in the role of chaos agents, trying to create more chaos) found that the other factions were more or less doing their job for them, making it easier to just amplify the crazy concepts others were coming up with. Again, that feels somewhat true to life.

I was at least somewhat surprised at the role that TikTok played in the various campaigns. Nearly all of the factions at one point or another came up with a TikTok strategy — perhaps foreshadowing where the technological battleground will be this year. Not surprisingly, much of the strategy of those supporting the Democrats in the election focused on first influencing what few swing voters remain, and then pivoted heavily towards getting out the vote and increasing voter participation. On the Republican side, there was a split as noted above. More traditional Republicans mostly ignored the Presidential campaign and focused on down ballot races concerning Congress, while the Trump campaign focused heavily on spreading fear, uncertainty, and doubt about… well… everything.

Running Threatcast has been quite eye-opening in highlighting the many different ways in which disinformation and misinformation is likely to show up in the next few months. If you’re interested in having us run Threatcast 2020 for your organization or group (it’s way, way, way better than a Zoom happy hour), please contact us.

Filed Under: disinformation, elections, games, misinformation, serious games, simulations, threatcast

I Helped Design The Election Simulation 'Parlor Game' Rebekah Mercer Got, And It's Not What You Think

from the the-strangest-timeline dept

How was your Monday? Mine was odd. Have you ever seen a news story where you know the reports are getting things wrong? This is like that, but on steroids. In fact, this was a story in which not only were the basic points wrong, but the wrongness started to go viral.

Jane Mayer, the New Yorker’s famed reporter, published a short piece talking about how Rebekah Mercer, the daughter of billionaire Robert Mercer, who helped fund Breitbart and Cambridge Analytica and helped get Donald Trump elected President, had “played” a new parlor game, called “Machine Learning President,” in which characters could role play the election:

In March, on a ski vacation at a rented house near Vail, Colorado, she brought a batch of copies of the ?Rules of Play? for an elaborate parlor game called the Machine Learning President. Essentially, it is a race to the Oval Office in three fifteen-minute rounds. It?s a role-playing game, more like Assassin than like Monopoly, although players of this game do start out with an allotment of ?cash? to spend on pushing their agendas, which can include ?algorithmic policing? and ?mass deportation.?

?Tonight, the name of the game is power,? reads the first page of the ?Rules of Play.? Each player, it goes on, ?will assume a new political identity.? Instead of becoming Colonel Mustard or Mrs. Peacock, as in the board game Clue, each player takes on the role of a political candidate or a ?faction,? in the game?s parlance. Among the possible roles are Mike Pence, Elizabeth Warren, Black Lives Matter, Russia, Y Combinator, Tom Steyer, Wall Street, Evangelicals, the Koch Network, and Robert Mercer himself.

There’s more in the New Yorker article, and you can read how Mayer describes the game, but, well, I was a part of the team that developed “Machine Learning President.” We didn’t design it for the Mercers to get to relive their successful electioneering… but that’s certainly the way the internet took the story.

Splinternews had a story entitled simply “What the Fuck???” and its sister publication Jezebel had one entitled “Evil Billionaires Play a Board Game Where They Pretend to Have Total Control Over the American Political System Ha Ha Ha Ha Ha”. Ha ha ha ha ha. Except, um, that’s not the actual story. Kate Conger at Gizmodo wrote up a much better version detailing the actual history of the game, but since I was there, I figured I might as well tell you the real story.

First off, no, this was not a game for rich evil billionaires to relive the election and the joys of playing with the populace. This was actually in response to the 2016 election, and how the role the internet played in that election took a lot of people — especially in Silicon Valley — by surprise. The founders of Scout.ai, Brett Horvath and Berit Anderson, thought it might be useful to do some scenario planning to game out how tech might impact future elections, as a way to think through the potential challenges and pitfalls that might eat away at democracy. The founders of Scout recently spun out Guardians.ai to deal with the problem directly and protect pro-democracy groups around the globe from information warfare and cyber attacks. The idea behind the Machine Learning President was to use the game to get people to think about these issues, and to prepare for potential abuses of the system, while also thinking through ways that technology could be used for good — to protect democracy.

Brett and Berit brought in game designer Randy Lubin of Diegetic Games, whose name you might recognize as our partner in releasing our adaptation of the CIA’s card game. Brett, Berit and Randy worked on the core of the game. Randy helped bring me in, along with science fiction writer Eliot Peper from Scout, to help think though various scenarios of how tech and politics might mix in the future. The five of us helped put together a fun game in which players take on the roles of various campaigns, special interests and factions, and run through a full election cycle. A key part of the game is how money and influence play into getting out a message and how technology and different potential alliances may factor into all of this. Again, the idea was a simulation to help people think through what might happen and how it might impact things. It was not to predict the future or to get people hyped up on destroying democracy.

Indeed, part of the goal of the game was to look for areas of agreement, or where there could be bipartisan or non-partisan approaches to things, where we might be able to uncover productive areas to fix major problems in a world where not everyone is at the throats of people on “the other side.” We had kind of hoped that the game might be useful in starting useful, nuanced conversations, rather than just people screaming at each other on Twitter. Ironically, the game resulted in people going nuts on Twitter.

The game was also designed to be non-partisan, and fairly balanced between the various positions and factions. It really is designed to be educational, and to see what happens. What kinds of technology may impact an election? What unexpected coalitions may form? How might certain candidates leverage technology or groups to achieve their goals?

We ran the game in San Francisco one evening in early February with about 40 attendees. It was a lot of fun, according to every attendee we spoke to. People got quite excited about it and certainly got deeply into the roles they were assigned. It was fascinating to see what kinds of deals people were trying to make, and how different groups decided to make use of their money or influence. It was not designed, nor could it be used, as a way to happily relive past electoral conquests. For what it’s worth, when we played in SF, the end result was Kamala Harris winning the 2020 election over Mike Pence, who defeated Condoleezza Rice in the GOP primary (don’t ask what happened to Trump) – in part because Rice, surprisingly, ended up endorsing Harris over Pence

And here’s the thing: it’s unlikely that Rebekah Mercer and her friends actually played the game, despite what Mayer’s reporting says (and what others are now reporting). From the reporting, it appears that she got her hands on the general rules that all the players got, as well as maybe a few of the character sheets. Each attendee played as a certain character, and each player got a sheet about that character, how much money and influence they had and what sorts of policies they supported as well as what overall goals they sought to achieve in the game. But playing the actual game requires a lot more — including a bunch of facilitators who understand the game, and a backend engine that they most certainly did not have (nor would they have understood it if they had seen it). Only the very small team of folks who were involved in the game have access to that.

Though, if anyone actually wants to set up an event to play this election simulation — feel free to reach out to us or to Scout and it can be set up. It actually is fun and educational.

Still, what was amazing was to watch people react online — almost all of them horrified, believing that the Mercers themselves had made a game that appeared to be mocking the public. The New Yorker itself claimed that it was a way for the Mercers to “relive” the 2016 election:

Max Temkin, the creator of Cards Against Humanity, unfortunately believed the Mercers themselves designed the game, and was pretty upset about it:

Others reported being “slack-jawed and stunned”:

Another popular one was for people to suggest that this all showed that the public was “playthings” and that this was “all a game” to them:


And my favorite may be reporter Ashley Feinberg who described it as “conservative billionaire dungeons and dragons.”

And, look, from Mayer’s initial reporting, I can totally see how people would jump to this conclusion. Lots of people don’t like the Mercers, and they appeared to have a game — and part of that game includes policy goals that many people find despicable, including mass deportation of undocumented individuals. And, of course, there’s a long history of governments, military, intelligence agencies, campaign and more using scenario planning games to anticipate the future. I mean, why do you think the CIA started creating games (that we’re now publishing…)?

But this game wasn’t designed to relive the last election or to celebrate these things. It was designed as a scenario planning tool to think how tech can impact future elections. And, again, the Mercers were most likely unable to play it without facilitators or access to the backend of the game. The game is interesting. But not in the way that people seem to have immediately responded.

Even more ironic: part of the reason for the game was to get people to think through what happens when disinformation and “fake news” about hot topics related to politics spread virally on sites like Twitter. And what happens? Misinformation about the game spread virally on Twitter. Life is weird sometimes.

Still, to some extent, we’re amazed that the Mercers found the rules of the game interesting enough to discuss or even attempt to play. It certainly suggests we created something that touches on important ideas. And the team of folks who created and ran that game are more than willing to host more rounds of the game in the future.

So, uh, anyone want to play?

Filed Under: elections, fake news, games, journalism, machine learning president, rebekah mercer, reporting, scenario planning, simulations

DailyDirt: Jocks Versus Nerds

from the urls-we-dig-up dept

The rise of fantasy sports and realistic video games for every major sport has expanded the audience and engagement incredibly. Even if you can’t throw a spiral, you can still manage a fantasy football team. Sabermetrics changed baseball, and deep learning algorithms are about to change how a lot of other sports are played. Computers aren’t just going to beat people at chess and Go. They might become better talent scouts and strategists for every major sport.

After you’ve finished checking out those links, take a look at our Daily Deals for cool gadgets and other awesome stuff.

Filed Under: ai, algorithms, analytics, artificial intelligence, basketball, cognitive computing, data, football, moneyball, sabermetrics, simulations, sports, toronto raptors, watson
Companies: ibm, nba, nfl