Blog / AI Capabilities Are Advancing Rapidly
Staying up-to-date about AI capabilities can help prevent nasty surprises.
Recently I encouraged organizations to consider addressing the use of artificial intelligence (AI) in their policies. At the time the capabilities of readily available AI systems were limited to text generation and answering questions. But of course we’re talking about computers, technology rapidly advances, and if you think you’ve got a handle on things just blink and something will have changed, including AI capabilities.
As the most charismatic, good looking, well-read, and most intelligent cyber-man out there, I don’t bring up my schooling very often so as to stay humble, but I actually graduated with a degree in computer programming specializing in artificial intelligence. The thing to remember about the current wave of AI products are that they’re not actually “intelligent”; they’re more like really good pattern matchers. After feeding an AI program a collection of data (the more the better), you let the software find relevant patterns in the data in the process you may have heard called ‘machine learning.’ This is why selecting your data is so incredibly important; the software looks for patterns on its own so it may find patterns you didn’t know were even there.
The thing is that true intelligence implies understanding as well as pattern matching, and new AI programs still don’t meet that criteria. For example, the sentence “Roses are red” is very simple, but to understand it you need to know both what a rose is as well as the colour red. Considering there’s no real understanding in ChatGPT or Bard or any of the other AI systems out there makes what they can accomplish that much more impressive.
So what’s the concern about AI capabilities?
The “big deal” is it’s not just chatbots anymore. Once developers demonstrated the technology works with human language, other quickly wondered “What about computer languages?” Almost overnight chatbots began “speaking” computer and generating working code. Of course, it rapidly became a tool for quickly developing or snippets of code that individuals would then fine tune. Even if they don’t “understand,” if you’ve used one you know these chatbots can be incredibly useful for quick programming and development.
Like all technology though, remember AI is a double-edged sword, and has already been pressed into service to develop malware as well. Before criminals either needed programming skills or at least access to someone willing to make your payload. A chatbot has no morals and will do whatever you ask it.
AI capabilities is a complicated topic but worth paying attention to as new developments in can impact your organization in new and unexpected ways, but not all of them good. If you have questions about AI capabilities and how they can impact you, contact one of our cybersecurity experts and get yourself some stress-free IT.
Todays quote comes from Henry IV, Part II: “Presume not that I am the thing I was.”