The use of AI for security solution or threat ?

Yesterday I had the honor to be present to an event hosted by UCD having topic " The use of artificial intelligence in cybersecurity. Having huge interest on artificial intelligence field, I didn't want to miss this opportunity to be present to such an event.

Representatives from many organizations have gathered at Moli Newman House to listen the speakers:

• Dr Marguerite Barry - UCD School of Information and Communication Studies
• Dr Giovanni Russo - UCD School of Electrical and Electronic Engineering
• Dr Mathieu Sinn - IBM  Research Staff Member -Cognitive Data Science
• Ian Trinder-  PwC  Director Cyber Security

I will try to make a short description of the main points of this topic. Machine learning and Artificial Intelligence capabilities are growing at unprecedented rate. These technologies start to be used in absolute all domains. There are not fields in  which AI are not implemented or will be implemented soon or later.  However AI models can possess some treats which I will present in the following moment:

1 Malicious use of AI

Building machine learning and AI models rely heavily on big data. These models are incredible accurate if the data they are using are correct. To understand how important is data I will give an example.  A certain bank decide to implement a machine learning automation system to try to prevent credit card fraud detection system.  A system to be reliable must contain accurate data. We can imagine what will happen if the data contains errors, the system will automatically give errors because the models are trained to make predictions based on that data.

2 AI Regulation

Is not yet know, how artificial intelligence will be regulated . Wisely will be that regulators should collaborate with researchers to prevent or investigate potential malicious use of AI.

3 Digital Security

One of the reason why automation systems are implemented or will be implemented is that make our job more easy, more fast and better. The example that I gave with credit card fraud detection is representative.

4 Fake News

We are continuously bombarded with fake news and propaganda via media channels . AI systems will be able in somehow to make distinction between a fake news and a good one. However the expansion of automated tasks can lead to expand threats associated with social manipulation. Authoritarian states can take huge advantages using these tools in order to manipulate their own people.