Artificial Intelligence & Trumponomics - A Brief Explanation Of Trump's Victory

For many decades, electoral mechanics were greased with a well-established metalworking fluid composed of three ingredients: two dominant political parties, voters and media.

For decades, TV, Radio and newspapers acted as powerful magnetic fields on voters. Their communication means always managed to align citizens in the direction of one of the government parties, pretty much like electrons in a particle accelerator.

Then each individual voter/electron started to become a transmitter on his/her own, through social media. Facebook, Twitter, Instagram, WhatsApp, YouTube, Snapchat have transformed the volter into an active emitter. No more "pensée unique", no more expert advice: everyone can now reach a global audience. At the same time, livecasts and videos have swept thinking and analyses. No more neurones, only emotions and hormones...

The appearance of Truth and fake news have replaced fact checking. It does not matter though, as long as the "likes" and "retweets" are shared as much as possible. The more a post or a video is shared, the more it becomes true. Digital crowds do not seek what is right, but what looks likely.

A bluff is a false but credible story. In poker, a bluff works as long as your opponent believes you have a better hand, so he can't call your bet or raise, conceding you the pot. The weaker player can counter this strategy by raising all-in fairly often, forcing all-or-nothing confrontations.

To understand why these dynamics are so crucial in politics, consider the work of John von Neumann, the incredibly gifted polymath who contributed to the Manhattan Project. Von Neumann loved poker because its strategy involves cunning, statistics, some luck too, but is never transparent. It also always depends on the counterstrategies deployed by opponents in real time.

Top Poker players misrepresent the strength of their hands, simulate irrational behavior, and deploy other mind games to confuse their opponents. In a simpler or more basic definition: they bluff. It was von Neumann's efforts to express bluffs in mathematical terms that helped him develop game theory, which has numerous real-world applications, cold war strategies foremost among them.

Oskar Morgenstern, Von Neumann's main collaborator, summarized the value of poker logic in 1961: "The Cold War is sometimes compared to a giant chess game. The analogy, however, is quite false, for while chess is a formidable game of almost unbelievable complexity, it lacks salient features of the political and military struggles with which it is compared." 

In other words, chess provides no opportunity to bluff (it's a game of complete information)…As Daniel Negreanu, the all-time strongest poker tournament player, put it to James McManus, a Bloomberg's reporter: "Trump's bluffs are very effective against (poker) level-one thinkers. His lies are so outlandish that people think they have to be true or he wouldn't have said it. But sharper players would pick him apart."

But in order to play a poker game on such a large number of voters, you need to have access to an immense quantity of voters data, and for each of them, in real time. Voters/emitters do not align anymore to traditional media, but their wordiness on social media becomes more and more predictable. Furthermore, this data allows to better anticipate votes from the silent population, people who did not yet make a decision, along with abstainers, who are all keys in all big elections.

The analysis of likes, posts or tweets allows a granular analysis of behaviors and the construction of very sensitive prediction models. Just as Libratus proved it lately when beating some of the world's best Poker players, you need AI to analyze behaviors and reactions to bluffs in real time.

Libratus engineers developed a technique called "nested endgame solving" which allows the bot to compute new strategies "in real time that best counters the humans' actions, while still guaranteeing that a strategy is balanced so the humans can't take advantage" - Noam Brown, PhD student with CMU who helped design the system. 

It has shown that AI is unbeatable to search and find patterns on considerable amounts of objects and parameters. Add to this the ability to analyze voters' reactions to Trump's tweets, as well as matching them with the pictures posted by such voters in their polling booth in real time, you can get a realistic insight on how to bluff and what to tweet...Companies such as Cambridge Analytica in the UK are pretty strong at this, or Peter Thiel's Palantir. Investigations are ongoing regarding the exact role Cambridge Analytica played in the US election. Donald Trump has chosen them, further to their involvement with the Brexit referendum, which he saw as a test market in terms of manipulation of the masses.

The turning point in Trump's victory happened in July 2016, about a month after Peter Thiel's backing. Could Cambridge Analytica and Palantir bring the needed data mining AI tools to Trump to collect such data and make sense of it, in real time?

While traditional media were analyzing and driving information using traditional polling techniques, Trump could as well play his favorite poker game, leveraging it with Palantir's AI, to win the presidential elections while everybody thought he would lose.

One of the things an AI system such as Libratus does well is bluff: AI is impervious to the psychological ups and down - “If you have $200 in the middle and $20,000 in your stack, you can bet that,” says Doug Polk, a poker pro who bested a previous AI built by CMU in 2015. “But humans don’t really like that. It feels like you’re risking a lot of money to win so little. The computer doesn’t have that psychology. It just looks at the best play.”

Palantir has provided largely secret assistance to the US Customs and Border Protection agency (CBP) in operating a system that tracks and assesses immigrants and other travelers, according to public records since 2012. Palantir’s list of customers includes the FBI, the CIA, the Defense Intelligence Agency, the Securities and Exchange Commission, Immigration and Customs Enforcement, the military’s Special Operations Command and other federal agencies.

Coincidentally, Thiel's backing up became official in June 07 2016. About a month before this backing, several documents leaked in the press, stating that more than 100 Palantir employees, including several prominent managers, have left the company last year. At that rate, the company was on track to turn over about 20% of its staff in 2016, almost double the average rate of the three previous years.

The loss of several key accounts such as Coca Cola, American Express or the Nasdaq may not be seen as a coincidence either, in Thiel's decision to support Donald Trump...Or Hillary Clinton in November 04 last year...Though the latter may have just been another bluff move...

Thank you for your attention.

Youcef DRIDI

Keywords: #CambridgeAnalytica #Palantir #RealDonaldTrump #Artificial Intelligence, #FakeNews #USElections