The development of technology has made human work easier and has also changed the election model in some countries and regions. Artificial intelligence, machine learning, and big data are being used to understand the unique psychological and behavioral characteristics of voters, create political bias, guide voters to vote for specific politicians/parties, suppress campaign opponents, and gradually dominate campaigns. Deepfakes and social bots are heavily used to manipulate voter opinions during election campaigns.
1. The impact of social media on elections
Big data helps political parties interpret voters' thoughts and behavior patterns, while social media and mobile apps are considered the main sources of big data. A large amount of user data is collected in apps such as Google Map, YouTube, X, Meta, Instagram, WhatsApp, etc. According to the survey, there were 900 million voters in India during the 2019 presidential election, of which about 500 million voters were mobile phone users with Meta and WhatsApp accounts, and 30% of them could be influenced by social media.
Today, political parties and politicians are increasingly reliant on new technologies that use big data, artificial intelligence, and machine learning to analyze voter psychology and behavior. During election season, big data is invaluable because it can be used to build targeted, customized campaigns. Political parties in democracies are also using big data to understand the profile of their constituencies, the households in their constituencies. Political parties and campaign advisers want to know who are first-time voters, who are mobile voters, how they behave, and what their family composition and economic conditions are. In India in particular, political parties and political marketing consultants are more interested in understanding the caste, religious makeup, and socio-economic status of voters in society, and this data can help design campaign strategies, select candidates, and form a pro-party narrative. Political parties and political advisers are using big data, artificial intelligence, and machine learning to conduct more sophisticated political communications and target voters in a more targeted manner.
2. Big data reveals voter preferences
Big data is increasingly in demand in order to understand voters' choices and concerns, predict election outcomes, and design campaigns. The most important means of collecting big data are social media and mobile apps. To gain more supporters, political parties have developed their own apps to connect with voters and collect big data to better understand voters' perspectives, behaviors, and psychological profiles, encouraging voters to participate more in campaigns. During the election campaign, political parties asked volunteers and supporters with the right to vote to download their app, and users were required to provide phone numbers that allowed the app to access the user's directory, photos, videos, messages, GPS, and Bluetooth data. During the 2008 U.S. presidential election, Obama released the iPhone campaign app. The app accesses contact lists from users' phones in different locations, and when they get that, the campaign encourages users to call their contacts on behalf of their political parties to canvass for votes and flag their attitudes.
All tiny data such as the user's contact list, sent text messages, and received GPS location history, etc., are collected, and this is because allowing the app to access personal information from the phone is a prerequisite for using the program. In the United States, smartphones have successfully replaced door-to-door propaganda campaigns. Obama's victory in the 2008 U.S. presidential election was considered a data-driven campaign. Obama's canvassing app provides a list of targets for volunteers, 1 million people have provided campaigns with Meta data of themselves and their friends, and Obama's email contact list has 20 million Americans. In the 2016 U.S. election, a British political consulting firm (Cambridge Analytica) obtained the personal information of tens of millions of Americans through a third-party Q&A app without the consent of Meta users, which was used to build a "psychological warfare tool" for Ted Analytica. The presidential campaigns of Cruz and Donald Trump built detailed voter profiles and, in the run-up to the election, campaigned Trump with provocative political content and ads targeting skeptical voters.
3. Use of the app in the election campaign
Apps lay the groundwork for a strong digital campaign strategy in elections. During the 2016 U.S. presidential election, campaign apps became more sophisticated. Hillary 2016 is Hillary's iOS app for the U.S. presidential campaign, where users can earn virtual and physical rewards for completing tasks.
In 2020, Phunware Software developed "Trump 2020," an app for Trump's presidential campaign, which is a gamification tool, news aggregator, and virtual event platform that offers donations, social media conversations, surveys, and more to maximize data collection and targeted advertising. MIT Technology Review argues that "data collection is perhaps the most powerful thing the Trump 2020 app can do." Signing up in the "Trump 2020" app requires the user's full name, phone number, email address, and zip code, and the app also makes a large number of permission requests for location data, phone identity information, and control the phone's Bluetooth capabilities. The campaign strategy helped about 40 million to 50 million citizens vote for Trump's re-election.
During the 2020 U.S. presidential election, Joseph Biden's campaign team launched the "Team Joe" app. The app is an organizational tool that allows users to text friends who support Biden and get updates on the campaign, with the primary goal of allowing communities and voters to use their networks to share experiences and their support for candidates. This relational organization leverages the influence and social connections of volunteers to give campaigns the opportunity to reach potential like-minded people. The app will let users know which of the user's friends and family the campaign team would like to talk to. Users can then text them directly to share campaign updates, ask questions, and gather thoughts on the 2020 campaign on behalf of the campaign.
In the app store, the "Trump 2020" app is classified as a news app, and in early October 2020, it had more than 2.6 million downloads, more than 15 times more than "Team Joe". The Biden app is classified as a social networking app and is suitable for users aged 4 and above. Trump and Biden have different approaches, with Trump's app targeting all supporters and having a longer list of permission requests than Biden's app, which is designed for only one purpose, "relational organization," where volunteers and organizers strategically reach out to friends and family to vote for Biden. If this massive, semi-automated "relational organization" changes, the scale will automatically change and form a new form of propaganda.
Narendra Modi is the official personal app of Indian Prime Minister Narendra Modi, which is used to disseminate news and campaign information to supporters. Users must register by submitting personal information such as email, phone number, and occupation before using this app. In March 2015, Narendra Modi was downloaded 10 million times, and the default access was to allow access to almost all information and data on the user's phone, including the user's contacts, photo albums, location, etc., which was used for election campaigns.
4. Deepfakes and Election Activities
Deepfake videos can show things that a person has never said or done. With deep learning, AI algorithms can tamper with actual video and audio to mislead viewers. Deepfakes have lowered the barrier to entry for non-professionals to make videos and are thriving as an inexpensive technical tool and application for producing fictional content.
One of the methods of making deepfake videos is encoding, decoding, and swapping faces. First, the AI encoder runs multiple mugshots of two people. The encoder then learns the similarity of the faces and compresses the image. Since the faces are different, one AI decoder selects the first person's face and the other decoder chooses the second person's face. Then, the encoded image is fed into the "wrong" decoder for face swapping. This process needs to be carried out on every frame in order to make it possible to produce a convincing video.
In recent years, deepfakes have gained global attention, targeting high-profile figures such as Barack Obama insulting US President Donald Trump and Mark Zuckerberg boasting about illegal access to data, but these are actually videos that have been created and synthesized through deepfakes.
The day before the election in Delhi, India, the chairman of the Delhi branch of the Bharatiya Janata Party (BJP), Manoj J. Manoj Tiwari addressed 15 million voters in English and Haryanwi through 5,800 WhatsApp groups, calling on citizens to vote for the BJP in two videos that went viral. According to digital media company Vice, the videos are all deepfake. The original video was about a completely different issue, with audio simulated and overlaid in English and Haryanwi, and lip tweaks to give the impression that Tiwari was actually speaking.
5. Social robots
Currently, social bots, with their trusted online profiles and advanced conversion skills, look like real users on the web, are already ubiquitous on social media.
Social bot behavior can generally be identified at three levels.
(1) Robot clustering, social robots often gather together and act randomly, so they are easier to identify;
(2) content level, where social bots tend to use emojis, exclamation points, or other content more regularly than humans use on social media;
(3) Activity: Social media bots are more active than human social media behavior.
Bots strategically amplify specific messages, manipulate the public, and influence the political process by amplifying propaganda and shaping popular narratives. Robots are programmed to post, like, and share information around the clock, and are heavily used for political purposes to influence public discourse, military conflicts, and election agendas. It only takes a few seconds for a social bot to disseminate a specific message. According to one study, X has 6% of bots that post 31% of bad information.
The study found that social bots play a strategic role in inciting Brexit polarization, with hashtags related to leaving decoupling dominating conversations on X, with less than 1% of accounts generating content accounting for one-third of all information.
In the U.S., nearly 400,000 X accounts (about 15% of the total U.S. X users) were social bot accounts during the 2016 presidential election, generating traffic to about 20% of online conversations. During the coronavirus lockdown in the United States, nearly half of the X accounts discussed reopening were social bots.
6. Prospects for the application of artificial intelligence and machine learning in election campaigns
The future of artificial intelligence and machine learning depends on the development of technology and the Internet around the world. Recently, in the election campaign in some countries and regions, people have noticed the impact of technological and Internet upgrades.
Technology has become easier to learn, with thousands of applications available for free online that can edit, merge, and even fake audio, video, and images. With an internet connection, these videos and audios can travel around the globe in a matter of seconds. Online campaigning is influencing voters' choices, and it won't be long before the low barriers to entry and low cost of technology threaten to move campaigns entirely online.
In the future, there will be more and more applications of artificial intelligence and machine learning in national and regional election campaigns. Custom software applications, deepfake videos, and social bots will largely replace door-to-door campaigns. Voter details will soon cease to be personal privacy, and the election management company will get all the details of voters through a custom-built app. In a word, the future of campaigns will be driven by data, artificial intelligence, and machine learning.
Disclaimer: This article is transferred from Meta Strategy, the original author is a senior researcher at Meta Strategy Think Tank. The content of the article is the original author's personal point of view, and this official account is compiled/reprinted only to share and convey different views, if you have any objections, please contact us!
Transferred from 丨 Yuan Strategy
Author丨Senior Researcher of Meta Strategic Think Tank
About the Institute
Founded in November 1985, the International Institute of Technology and Economics (IITE) is a non-profit research institute affiliated to the Development Research Center of the State Council, whose main functions are to study major policy, strategic and forward-looking issues in the economic, scientific and technological and social development of the mainland, track and analyze the development trend of the world's science and technology and economic development, and provide decision-making consulting services for the central government and relevant ministries and commissions. The "Global Technology Map" is the official WeChat account of the International Institute of Technology and Economics, which is dedicated to conveying cutting-edge technology information and technological innovation insights to the public.
Address: Block A, Building 20, Xiaonanzhuang, Haidian District, Beijing
Phone: 010-82635522
WeChat: iite_er