It is imperative to understand that polarisation is a major driving force for the vulnerabilities in a democratic system. With a looming descent to a new Cold War between the west and the east, the information space is a primary front of war. Secret agents used to need to get to the top politicians and dig up dirt on their lives – to an extent this is still happening of course, scandals and kompromat are as present as ever – but AI is becoming much more cost-efficient route to get to the politicians: Influence through their voters. Malicious state actors have been provided ample tools by the western social media corporations to simply buy their way into the hearts and minds of the people. And if we are completely honest, Russia and China in the ages of communism, practically wrote the book on manipulating public opinion on the home front.
With the advent of social media and surveillance capitalism, the same tricks have become massively applicable in our media environment. Journalistic institutions built to combat the soviet misinformation machine no longer carry weight in our daily political discourse or direct the media away from taking the bait. We have brute-forced an idealistic representation of democratic media representation, ignoring the foundations of what made our western media so potent in repelling information attacks from across the borders in the last Cold War.
As we idealistically plowed forward, we ended up producing a fertile ground for misinformation and propaganda from malicious state actors like Russia, Iran, North Korea and China. Now their efforts can be hidden and the volume of the broadcast be too high to meaningfully filter: “flooding the zone with shit”. Media literacy can help, but the problem is much deeper; There are fundamental vulnerabilities in human psychology that have already been exploited by the algorithms. These won’t go away precisely because they are fundamental. And for profit margins of the social media corporations to be upheld, it is necessary for them to keep exploiting these vulnerabilities that provide the attack vector for propaganda to begin with.
It used to be the job of journalistic oversight institutions to comb through media and recognise where outside influence and foul play was being engaged in. But now the task at hand has been made impossible.
For social media companies to not exploit the psychological vulnerabilities of humans would require them to actually staff humans with the capacity to understand complex moral questions in politics and positions, where they now simply put the computer with meagre capacities to understand these problems. And even with the capabilities of large language models improving, it is unlikely that they will improve enough in a short enough span of time to actually do anything about the problem until it is too late. Even if such a miraculous breakthrough in artificial intelligence actually is achieved, the computing power will still be too expensive to make it impractical for the companies to deploy such systems on the massive scale required for a misinformation & propaganda-proof social media.
Toni Aittoniemi
Pirate Party Finland
Gimulnautti 