Artificial intelligence goes to war

Introduction

Artificial intelligence goes to war. This is the headline that comes to mind when we see the news we receive from the war between Russia and Ukraine. There are countless fields in which artificial intelligence is beginning to be essential, and the military field was not going to be any less. Already in 2017, the Russian president, Vladimir Putin, stated: «Whoever is the leader of artificial intelligence will dominate the world». And he was not wrong, countries like the US, Russia, China or England are making great efforts to strengthen themselves in this field since it can really be decisive in the future. We will have disinformation campaigns on social networks, creation of bots, data analysis, identification of objects and people... The war of the future is already here.

The war of the future

In Ukraine we are seeing how activities that will be key in the war of the future are being tested. The US is using this war as a testing ground. These tools today are processing countless data from the Russian military. This will serve to predict future steps in Russia's strategy. According to Maynard Holiday, director of defense research and modernization engineering at the United States Department of Defense, the tools developed by the North American country will be capable of changing the course of any war conflict in the future. Conflicts in which decisions made by humans will take a back seat, making way for software whose decisions, a priori, will be more efficient.

American aircraft project of the future

Applications

It is well known that artificial intelligence can make decisions much more efficient and faster than humans, which is why different tools have begun to be used:

OSINT, the big brother

We will call him Big Brother because he is the all-seeing eye. OSINT, acronym for Open Source Intelligence, is a tool whose mission is to make use of public sources to be able to search and collect all the information. This, in a war context, translates into analyzing all those sources of information that we know to generate attack, infiltration, defense strategies... The sources of information are: media, social networks such as Twitter, Facebook, Instagram...

Clearview AI, the facial recognizer

Ukraine has begun to use this facial recognizer with the aim of facilitating the task of identifying Russian enemies. The American company, clearview, has millions of images obtained from VKontakte, the Russian social network similar to Facebook. It is said that all these images are being used by the Ukrainian army to detect enemies.

What could go wrong?

We will agree that using artificial intelligence in a field like war, in which every decision counts, should at first seem like a great advance. But it's also true that their algorithms could be incredibly easy to trick into seeing things that don't exist, and vice versa. So anyone who learned to confuse their rivals' systems would gain a huge advantage over others. We have a clear example of one of the Achilles heels that AI could have in an experiment carried out by MIT researchers. In this experiment they got Google's AI to mistake a turtle for a rifle. They achieved this thanks to the modification of the object in a 3D version, and although it is easily recognizable to the naked eye (human), the AI ​​found with 90% reliability that it was not a turtle but a rifle.

This is where we find a clear Achilles heel, the MIT researchers developed an algorithm capable of reliably fooling a machine learning model. That is, enemies could be deceived by camouflaging tanks, planes, military personnel, etc. so that they would not be recognized by their systems.

Ethics

The US government had a project whose objective was to automate the identification of aerial objects in images. The American company "Google" was contracted for this project. The company abandoned the project due to protests by its workers. After these internal protests at the multinational, they created a new ethical code for the use of AI. In it they tell us that They will not use their algorithms to develop weapons, or other technologies whose primary purpose or implementation is to directly cause or facilitate injury to people. That is why a debate is opened, Should we use artificial intelligence to cause harm?, Do we really want artificial intelligence to go to war?

Bibliography

Altim (April 2022) https://www.altim.es/blog-noticias-tic/la-inteligencia-artificial-un-arma-mas-en-la-guerra-de-ucrania/

MIT (October 2019) https://www.technologyreview.es/s/11563/la-paradoja-de-la-ia-militar-un-arma-peligrosa-que-nadie-debe-obviar

MIT (June 2018) https://www.technologyreview.es/s/10297/los-siete-principios-de-google-para-evitar-criticas-sus-proyectos-militares

Antena 3 (November 2017) https://www.antena3.com/noticias/tecnologia/google-confunde-tortuga-rifle_2017110559fee7490cf2ebaa1659924d.html

El Confidencial (April 2022) https://www.elconfidencial.com/tecnologia/novaceno/2022-04-25/pentagono-rusia-ucrania-inteligencia-artificial_3413944/

Nationalinterest (March 2022) https://nationalinterest.org/blog/techland-when-great-power-competition-meets-digital-world/artificial-intelligence-and-future

Written by

Alejandro Ferrer

More artificial intelligence!

The mission of Saturdays.ai is to make artificial intelligence more accessible (#ai4all) through intensive courses and programs where projects for good are carried out (#ai4good).

Find out about our artificial intelligence master's degree in https://saturdays.ai/master-ia-online/

If you want to learn more about artificial intelligence join our community at community.saturdays.ai or visit us on our website www.saturdays.ai Join us!