Kushay's Matter Bank Politics Science and Technology

[AK] How Artificial Intelligence Can Be Used To Influence Politics

This note will discuss how a set of artificial intelligence tools can be used to sway public opinion and attract them to a certain idea and the harms it brings to democracy.

A tech company specializing in big data and artificial intelligence named Cambridge Analytica is now the biggest of its kind. It is owned by Trump supporters like the Mercer family and Steve Bannon also serves as its board’s member. It has successfully swayed public opinion to put Trump in the White House, help increase support for his policy, etc. How does it work?

First, the machine obtains big data from various places like your Facebook likes, the YouTube videos you watch, the TV programs you watch, etc. and develops a model to predict your personality from it by matching the data with an “OCEAN Personality Model” (look it up :)). The effectivity is staggering – Michal Kowinski, the one who develops his model, claims that 10 data about Facebook likes enables the machine to know him better than his coworker. With 70, it could ‘know’ a subject better than a friend; with 150 likes, better than their parents. With 300 likes, Kowinski’s machine could predict a subject’s behavior better than their partner.

If the employer of Cambridge Analytica (or similar corporations) are already in office, the data can also be obtained from government agencies like NSA, Homeland Security, etc.

Second is targeted marketing: Since INDIVIDUAL voter’s personalities and issues of concern has already been mapped out, personalized message designed to change the people’s personality to support a certain policy that is super specific to the voter’s personality and concerns is spread out. The message comes in various forms: Facebook advertisements, dark posts (posts that only the recipient can see), fake news, emails, TV advertisements, bots and fake accounts, and so on. But notice that it isn’t the policies that is adjusted to people’s personality but rather, it is the message used to persuade people to support an already determined policy. For example, if a politician wants to garner support for a policy to oblige people to show their birth certificate on voting day: If the subject has a personality that doesn’t like commitments, the message might be “you should support this because it’s super easy to do”. But if the subject has a personality that is more Stoic traditionalist, the message might be “you should support this because it is your obligation as a democratic citizen”.

Lastly, the subject’s response (whether he clicks on the link or not, his comments, what kind of “like” emoji he uses, etc.) is recorded and used again to provide message that is increasingly adjusted to his personality.
The results are staggering. Two examples:

1: Ted Cruz, who initially was only known by below 40% Americans (as opposed to Jeb Bush’s 85%) rose rapidly to become the strongest contender against Donald Trump.

2: Margin shifts in traditionally blue states (states where democrats win was sufficient to make it a red state (states where republicans win) in Pennsylvania, Michigan, and Wisconsin. These three states won Donald Trump the presidency. While in the election campaigns, Hillary’s campaign team (who mostly uses traditional pollers instead of AI) laughs at Donald Trump’s decision to focus his campaign in those three states.

The implication to this is very terrifying. It means that democracy will shift from a constructive discussion of opposing ideas to a game of changing people’s personality via exploiting insecurities and concern. Moreover, assuming that both republicans and democrats uses this technology, the likelihood of incidents being artificialized is bigger (for example: An artificialized mass shooting to exploit people’s fear of gun ownership) since there is virtually no incentive to approach the voters on a logical level.

Outside democratic countries, the effect is even more scary. If countries like China or Russia (who are already autocratic) uses this technology, then they literally control the country regardless of how messed up their citizens is. Ever heard of North Korea? The people there are poor, famine is widespread, yet the society there supports their government and demonizes the west instead because of the massive propaganda done by the regime.

A possible solution to this menace might be a group of counter hackers that also moves on a same scale, but points out to the voters that some of the messages they receive are artificialized.

Leave a Reply

Your email address will not be published. Required fields are marked *