Artificial Intelligence and Valentines Day
February 14, 2020
As I put the finishing touches on this article, it is Valentine’s Day. When I think about Artificial Intelligence (AI) in the context of Valentine’s Day, and especially human emotions - such as love! - I wonder what our future holds. AI as a research discipline has made incredible strides since 1956, but truthfully not always delivered on its headier promises. This has led to AI being considered a trendy buzzword without real-world applications. Yet, examples abound of AI beating world champion humans at their own games (like chess, Go, or Jeopardy!), sifting through unimaginable amounts of data, creating deep fake voiceprints/videos, and outperforming radiologists on breast cancer detection!
But let’s step back. What exactly is AI? In short: Any environment-aware device which acts to maximize its own goals. Popular notion of AI includes “intelligent agents which mimic human mental functions such as learning or problem solving.” Common with other technology fields, there are many approaches to AI (such as Weak AI, Strong AI, and superintelligence), and it has dozens of sub-disciplines, hundreds of applications, and many design elements.
How do we use AI in Information Security?
Most of our AI use is in systems which take large volumes of logs or data and make sense of what is legitimate and what needs to be investigated. Our Security Incident Event Management (SIEM) dashboards and filters alert us to anomalies within the millions of logins, network connections, emails, etc. for our Security Operations Centre (SOC) staff to investigate. The SIEM also ‘learns’ in the sense that it shows us daily trends e.g. how this Monday compares to the previous four Mondays. It stands to reason that the trend for the first Monday of classes is higher than the baseline, but not the other way around. We also have several protection systems which first operate in “learning” mode before we use them to automatically block email or network connections. So for us, AI truly is a helper to automate what are often “one or more needles in many haystacks” problems. In this way, they help us serve and protect you.
What about the future of AI?
People often ask: Will computers ever be smarter than humans? Well: Yes, and No :-) It depends how you define it. Or: Can computers have emotions like us? Again, it depends on how you define it. Driverless cars will need to perform appraisals to make ethical decisions, but human emotions are a combination of psychological and biological inputs. So, Canadian professor Paul Thagard says – with many others: “Unlikely”. But what if we fast-forward a few decades and imagine tiny AI devices implanted through our body which intercept and ‘read’ e.g. the ~100 million inputs from the digestive tract alone. Will we have robots which care about “their” humans? Will we have devices which, in addition to true “Information”, “Data”, and “Knowledge”, also have true “Wisdom”? (Cf. the DIKW pyramid.) I personally do not think so. I believe we will keep re-defining what it means to be human so that devices do not take over humanity. So even when robots build robots, I believe humanity, with all its imperfections, will remain central. In fact, the alternative is unacceptable to me personally. A dating algorithm may help me, but the feelings of love are up to me. Happy Valentine’s Day!
The field of AI is vast and fast moving. It was a real challenge to pick out just those interesting items which make this a suitable blog post, as opposed to an in-depth article or a monograph. So if you are interested in further reading, here are some selected links.
Also: search for Artificial Intelligence in the Gartner Database at: https://guides.lib.uoguelph.ca/az.php?a=g
Written by: Gerit Bos, Information Security Officer.
Image credit: Composite created by Gerrit Bos from source images courtesy of pixabay.com