Introduction to Security in Artificial Intelligence
Welcome in my first post. I will be writing about an IT issue that is Artificial Intelligence, and my research will focus on Security Industry. Why would I focus on that topic in particular? Many of us have no idea how much AI is being used around us to watch for threats and to observe us, so let us dabble into this hidden world.
I will be proceeding with basic and survey research as I think it is the most appropriate method for the above topic.
In the year 2018, when working as a Security Officer in a store in Ireland, I have been asked by my manager to look for a specific person on CCTV, that entered the store a day before at an unspecified time and follow his movements around the store to see if he might have stolen anything. To find when that person entered and to follow him on over 15 different cameras took me around 5 hours, mostly due to how slow the computer worked while shuffling different camera angles. October 2024, when entering a shop in Leicester (United Kingdom), I had a glance at the camera screen above the entry, and behold! My face with time and date of entry, in such sharp quality I could see my moustache from up close. Assuming the pictures are saved for a while, it would safe my previous boss hours of my salary to find the same results.
As you yourself can learn a body language of someone that is about to steal something or perform an act of violence, is there any reason an AI system could not be thought the same?
One of these examples might be development of "Hawk-Eye", a threat detecting program installed in smart cameras. While not resource heavy, this AI is designed to recognise potential threats and send a signal to human security team. It can detect a variety of weapons or gun arms, with precision to up to 99.99%. In extreme situations, it might prevent mass shootings or even terrorist attacks.
Face recognition such as that might be also used in the airports, improving detection of potentially dangerous or wanted people, making the experience of going through the entire airport experience much more pleasant. At least for passengers that are currently not wanted by local or international forces.
One such program already has been tested since in years 2016 to 2019 in Hungary, Romania and Greece, called "iBorderCtrl". Main difference is it focused on land border crossings. An electronic avatar performed questioning any normal border patrol officer would, detecting lies, comparing face to the passport photo or asking for the reason to visit. If anything suspicious has been detected, you were sent to the human operated part for further questioning.
_from_U.S._(Blaine,_WA).jpg)
Programs as such were a topic of controversies since the projects inception in 2016. Common worries consisted of excessive collection of peoples data, lack of transparency about the testing of the program, lie detector not working properly with people with anxiety or disabilities or creating peoples online profiles and categorizing them. How much freedom are we ready to surrender for what people in power call safety? I would love to believe the government has our best interest in mind, but it is probably my wishful thinking.
While most of us live our lives in relative peace, our armies are always ready to mobilize in case of a national threat. Since the dawn of time and human conflict, weapons had evolved to maximize damages on the opposing side while minimizing them on initiators side.
On a more positive note, another example of everyday artificial intelligence that is being by used is scam and phishing detection. Phishing can be described as an attack to try and steal other peoples passwords or personal data. The oldest example I can remember were e-mails from a Nigerian prince trying to transfer his wealth to you. In recent years phishing and online scams has been steadily evolving. As many computer illiterate people have become victims to scam calls, Samsung and McAfee companies had developed an AI that detects scam messages and flag suspicious calls, and I personally witnessed it working in my Samsung phone before.
This has been my introduction to the topic of AI in security. In my next post we will be going into my primary research, so stay tuned and I will see you there.
Sources:
Ahmed, A.A. and Echi, M. (2021) Hawk-Eye: An AI-powered threat detector for intelligent surveillance cameras | IEEE Journals & Magazine | IEEE Xplore, https://ieeexplore.ieee.org/. Available at: https://ieeexplore.ieee.org/abstract/document/9408578 (Accessed: 25 October 2024).
CORDIS, cordis. europa. eu (2016) Project flyer: H2020: Cordis: European Commission, CORDIS. Available at: https://cordis.europa.eu/project/id/700626/reporting (Accessed: 25 October 2024).
What is iborderctrl? (no date) iBorderCtrl? No! | iBorderCtrl.no. Available at: https://iborderctrl.no/start (Accessed: 25 October 2024).
Breyer, P. (2021) EU-funded technology violates Fundamental Rights, aboutintel.eu. Available at: https://aboutintel.eu/transparency-lawsuit-iborderctrl/ (Accessed: 25 October 2024).
O’Connell, R.L. (1990) Of arms and men: A History of War, Arms and Aggression. Oxford: Oxford University Press.
Iwan-Sojka, D. (2024) Responsibility of war industry for Lethal Autonomous Weapons Systems, Wydawnictwo Uniwersytetu Śląskiego. Available at: https://monograph.us.edu.pl/index.php/wydawnictwo/catalog/book/PN.4209 (Accessed: 28 October 2024).
Phishing (no date) Phishing | Information Security Office. Available at: https://security.berkeley.edu/education-awareness/phishing#:~:text=Overview,spam%20from%20compromised%20email%20accounts (Accessed: 28 October 2024).
Comments
Post a Comment