University of Tampere’s Evil-AI-project researches the dark side of AI.

University of Tampere’s Evil-AI-project researches the dark side of AI. 

AI and robots are dual-use tools. 

Evill-Ai-Project researches issues like the misuse of AI. Where there are computer systems. There are always vulnerable systems. Even the best systems are somehow vulnerable. And the thing that can cause a collapse in data security. It is one paper that the professor forgot on the wrong table. The hacker can see screens from outside. And the opportunity makes thieves. The AI is coming to homes, and the regular merchandise can turn against us. Wireless headphones and a loudspeaker-microphone combination can be used as eavesdropping tools. 

The intelligent TVs and smartphones can collect information that makes us vulnerable. And the robot vacuum cleaners can observe us. Those things are really bad. Researchers. Highjacked a vacuum cleaner and made it ask for credit card numbers. Another test was that. The researchers used an AI algorithm on a robot vacuum cleaner to hunt a cat. The fact is that this kind of malware in robot cars can turn lethal. The same robot that transports food to humans can transport explosives. 

The Soviet, India-class submarine with two midget submarines. Those systems can carry even nuclear detonators internally or externally. In the last case, the flat nuclear device can be on the shelf, and a robot hand can put it. In the right position. Then, sonar or some other system sends the detonation signal to that system. The same AI. Used for food delivery, robots can be used in those midget submarines. 

Or the land drone can carry machine guns and anti-tank missiles. This kind of system can also drive on the seafloor. They could be used as reconnaissance or attack tools. Sometimes it is suspected that those underwater tanks, operated by SPETSNAZ, left nuclear detonators on the seafloor. The remotely operated nuclear mine, with 1-15 kt. could cause very bad damage in harbours where it detonated. The nuclear device might look. like an anti-tank mine. 

The idea is similar to systems that we can call “underwater tanks”. Those systems are submarines, equipped with caterpillar tracks. This type of miniature submarine can have a ball-shaped instrumental pack. Which. Allows it operate very deeply. They can also transport nuclear explosives. 

Another thing is that the robot vacuum cleaner can transport even landmines under vehicles. And this means. The AI system is easier to misuse than thought. If we think about the possibility. To create the attack tools. By using. The AI, the same algorithm that allows the AI to hunt cats around the house, makes it possible to use similar algorithms to control the flying quadcopters and other kamikaze drones. This makes those systems dual-use. 

https://evil-ai.eu/about/

https://www.tuni.fi/en/news/jane-and-aatos-erkko-foundation-grants-eu1-4-million-advance-expertise-managing-dark-side-ai

https://yle.fi/a/74-20190991

https://en.wikipedia.org/wiki/India-class_submarine

https://en.wikipedia.org/wiki/Spetsnaz

Society without security destroys itself.

We always ask. What is the biggest threat to our society? We can answer that society itself is the biggest threat to itself. We say that surveillance brings security. But we forget. That in the wrong hands, surveillance can be an ultimate tool to control and destruction. Security is the balancing of freedom and privacy. Privacy is always important. But some people, like terrorists, misuse privacy. 

If authorities react only when things like robberies are done. That might cause terrible situations. Surveillance destroys privacy. And it's a very devastating tool in the wrong hands. The problem is in cases where surveillance operators. Hide information for blackmailing. 

Or they found something from the “grey area”. A grey area means. Something that is against the moral code. But they are not illegal. Those things might be somehow against the social code. And they can be embarrassing. Those things. Can cause. Bullying or social isolation. 

Society must be prepared for internal and external threats. 

The biggest threat is a collapse that causes internal disorder, or some other situation that society cannot resist, whether from internal or external threats. The disorder opens a road. For outside actors to take the country. Under control. The disorder creates a situation where the authorities cannot protect the people. And this is. The major problem for protective authorities. Who protects us from protectors, who turn against us? The AI-based surveillance systems threaten our privacy. At the same time. 

Security personnel are working with new types and more complex threats. There are lots of people. Like in places like subways. The people who look at surveillance camera screens. Must find things like bombs. They also must find missing kids, people who might look like suspected pedophiles, and drug dealers. And people who might be dangerous. There are billions of things that can go wrong. 

And in that case. The chaos can turn very dangerous. The society must protect its members. When somebody steals something from shops. There must be something. That confirms testimonies. This is the key problem in every type of criminal investigation. Things. The people in the court must be confirmed. When somebody steals from the market. Stuff. That they took. Must get back. If the crime is done. It’s hard. To get stolen property back. If somebody is injured or dead, nothing will bring that back. 

The problem is how to find a person on some streets? There is a possibility that police stop every single person and search for things like used weapons and stolen property. This is possible if there are 20 people in a village. But if there are 2,5 million people in some area. This takes a very long time. Things like AI-based cameras and terahertz scanners are made. To remove the need for unnecessary body inspections. 

This is one of the biggest problems. If guards or police ask who stole something. The person whom they ask. Can. A self is a thief. That person can tell lies. The fact is that. There are false alarms. But there must be something. That confirms that an outsider steals merchandise. It’s always possible that guards or staff steal something. 

But another thing is that. The same systems that can recognize people who steal, or who are nervous. Detection of dangerous people. Based on the behavioral model. People who plan something illegal. They are nervous. The problem is: how to determine “illegal”. Is it some kind of robbery? Or is just talk against the ruler enough to make something illegal? And then. We must realize that there are limits. Also for free speech. Things like hate talk, trolls, and threats are things that require a reaction. 

Those systems that detect nervous people. Can also be dangerous tools in the wrong hands. But everything can turn dangerous in the wrong hands. People can turn almost everything. Into eavesdropping systems. Then we can think. About telephones and their location capacity. Those telephones can. Tell the location of the tracked person.