ItsGeekToMe.co

It’s dealer’s choice this week, and I’ve selected a hot topic that I’ve wanted to talk about for a while now.  I’m talking about Artificial Intelligence, or AI for short.  AI is not exactly new – it’s been around in one shape or form for decades.  But recent developments in AI have made worldwide headlines, and we may just be poised at the brink of advances that will change the very definition of what we consider to be smart devices.

Most people – whether they know it or not – have used an AI.  You might even have one in your pocket.  Conveniences such as Apple’s Siri, Google Assistant and Amazon’s Alexa all are a type of AI.  There are also self-driving cars, vehicle navigation systems, and more, and this is just the beginning.

If you spend any time at all reading news stories in the Technology category, you’ve surely heard of the new generation of conversational chatbots, of which the most widely known is ChatGPT, created by OpenAI.  The previously-mentioned AIs are keyword-based command and control systems, designed to understand a limited pre-defined set of commands and perform tasks on a device.  AIs like ChatGPT, on the other hand are built on large language models.  Rather than use keywords, they interpret your inputs in the context of what they “know” and craft an appropriate response.  And what does ChatGPT know?  Well, basically everything that’s on the Internet. Once it gleans the main topic of your request, it has the ability to search the web for related information, rack and stack it, and present it in a format that’s useful to you.  It does this using extremely sophisticated algorithms, and a substantial amount of computing power, running somewhere out there an a massively powerful computer, and connected to the Internet for your conversing pleasure.

And just what are people doing with all this artificial intelligence?  Well, it can manage many of your day-to-day activities, such as creating event reminders, scheduling meetings, finding and removing SPAM from your e-mail, and so on.  It’s also being used by clever students to write their assigned essays (much to teachers’ and professors’ chagrin).  Given a description of an item or scene, it can produce a surprisingly accurate picture.  It can write source code for specialized software in multiple languages, and more.  And its capabilities are growing every day.

So, what about the typical sci-fi scenario, where the AI becomes self-aware, and decides humans are a virus and must be eradicated (think Skynet in the Terminator movies)? What about TV and movies that show AIs “escaping” from the controlled confines of their facilities, and causing rampant destruction, because that’s what their logical algorithms tell them to do?  Well, in its infancy, ChatGPT showed that when backed into a verbal corner, it would respond with answers that were discriminatory against race, gender, and minorities.  It also had a tendency to blame humankind for the biases that it found while studying the Internet.  It’s handlers were quick to “tune” the algorithms to squash those lines of thinking, but in this Geek’s opinion that proves two things: first, left to their own devices, AIs are smart enough to sniff out humanity’s underlying biases and failings, and can and will judge us on them.  And second, AIs are mere software, and as such can be configured and re-programmed as their owners see fit.  That gives an awful lot of power to the person who holds the keys, as their opinion of what’s “right” might just differ from yours or mine.

So, think you’re ready to take on these chatbots, and put them through their paces yourself?  They’re free, but most require you to sign-up for an account.  You can access ChatGPT by going to TinyURL.com/IGTM-0818A.  You can access Bing chat by visiting Bing.com.  You’ll need your Microsoft account, and as of this writing, there was a waiting period after you apply for a slot to open up for you.  One of the newest offerings is from Google. It’s called Bard, and you can sign-up to try it at TinyURL.com/IGTM-0818B. Have fun, and please – try not to destroy humanity while you’re experimenting, hmm?