AI for Bad? Artificial intelligence and Disinformation

Science & Tech

Artificial Intelligence (AI) is increasingly affecting global security. In addition to its influence on the global economy and its impact on international relations, the AI revolution is increasingly used as a political tool capable of shaping public opinions.

Artificial Intelligence (AI) and machine learning are strong technology enablers, promising significant positive impact on people’s lives by making services easier and more efficient. However, the growing imbalance between the power of big tech companies and their users and the possibility to make decisions without human intervention give rise increasingly to concerns and criticisms among the population. At many different levels, the AI revolution brings threatening and potentially harmful effects for global security. One particular challenge is AI’s role in facilitating propaganda and disinformation. Developments in this field and their effects on democratic societies warrant scrutiny.

The Information Ecosystem
AI is playing an increasingly important role in the information ecosystem by enabling propaganda and disinformation to be more efficient, scalable, and widespread. AI mechanisms enable tailoring propaganda to a targeted audience and increase dissemination on a large scale. Highly precise voter profiling, the control of users’ news environment and online messaging, the promotion of biased content through a limited number of media platforms, as well as disinformation targeting public sector employees to influence government operations are only a few examples of what AI enables.

Behavioral data collection and exploitation
Every day, vast amounts of data are being generated by people using internet and social media.By using machine learning, it is possible to analyze this data, build profiles, and even to determine precise consumer behaviors. These mechanisms are first and foremost used for advertising purposes. However, they are also being used for political propaganda. Machine learning algorithms on social media enable a granular profiling and the possibility to send targeted political ads based on the profiles of voters, exposing them primarily to content they are susceptible to agree with. The audience is then filled with a tailor-made set of opinions, which reinforces their pre-existing positions and prevents them from an exposure to a more diverse set of ideas. The typical case of Cambridge Analytica’s Psychographics, famous for its implication in the 2016 American presidential election, is an enlightening illustration on how AI mechanisms can affect electoral processes by leveraging Facebook data and use personality assessments to tailor messages and content to individual users. Besides, we know that half of the online traffic is made by bots—web robots programmed to do certain repetitive tasks such as scanning content or chatting with users. These bots, which are imitating human users’ behavior, can be refined through AI as “political bots” working for a specific candidate, or party, through multiple actions such as retweeting specific messages or attacking other political candidates by spreading defamatory content. These actions affect political discussions without the general public necessarily being even aware of it.

Shadowing Tomorrow’s New World Order
Consequently, using AI mechanisms, both private and public actors can shape information flows and can affect democratic processes and even mold global public discourse and opinion. The nature of AI tools enables a small number of people, whether from government, corporations or other interest groups, to quickly get a substantial political influence and causes significant threats of for democratic systems and international security.
Even though some forms of machine learning can help detect and filter propaganda content, there are no appropriate tools to prevent and control the threats enabled by AI. For now at least, the latter outperforms existing preventive mechanisms.

From a global security perspective, it is fundamental to understand how innovations will shape the global economy, the media and societies, but also how AI technologies are already shaping public opinion. World leaders of tomorrow will be the ones with access to the best data, computing and human resources. New mechanisms of regulation, as well as trust and confidence-building measures need to be put in place. Nevertheless, it seems today unlikely that AI and its effects could get the same international control architecture as nuclear arsenals, for example. In any case, such control tools are yet to be invented, and they may be as hard for us to imagine as internet was for the pre-1990 population. As Putin says, “the one who becomes the leader in AI will be the ruler of the world”, for better or for worse. Probably for worse.




Image retrieved from dreamstime