Chatbots and Artificial Intelligence

[Marjo]

Hi guys,

Following our earlier discussion around digital home assistants, I'd love to see some discussion around another interesting type of automation...

Chatbots and Artificial Intellingence. 🤖

Do you generally understand the difference between bots, chatbots, and AI? Here's a quick summary:

• Bots are simple automated tools, used to make tasks easier; they can only recognise keywords and execute actions based on them. They present you with a list of buttons that you can click to move the interaction along to the next stage.

• Unlike the above, chatbots can actually interact, but still rely on human programmers. Chatbots have natural language understanding, which means they can understand a broader range of language and interact, to an extent, by asking questions, interpret responses and execute actions. What separates them from AI is that they still require input of programmers to function.

• AI (Artificial Intelligence) can learn from its interactions. An AI uses deep learning to receive data and interpret it, and these interpretations are not determined by programmers. Gartner defines AI as technology that: “appears to emulate human performance typically by learning, coming to its own conclusions [and by] appearing to understand complex content”. It also engages in natural dialogue with people.

Do you know of any examples of the above?

Have you used them? How was the experience?

What situation do you think bots / chatbots / AI would be good for?

Anything you feel you would not use a bot for?

[Glory1]

Now reading that article, @jonsie, really is terrifying and goes somewhat to what I was trying to say. Yes, you can create a super intelligent AI but should we?

Hate to say it but the film Terminator comes to mind. A piece of fantasy, yes but is it?

[pgn]

Off to read this, on the strength of the link @jonsie  gave...

https://thenextweb.com/google/2019/05/16/googles-new-ai-can-help-you-speak-another-language-in-your-...

and more detailed here:

https://ai.googleblog.com/2019/05/introducing-translatotron-end-to-end.html

Skynet next?

[pgn]

It has taken me a while to read up on the AI scene, and I watch with interest, hoping we never get to the likes of Minority Report (not too far from some of the news lately wrt to facial recognition cameras!) and similar.

I Robot's 3 laws, (Asimov's, I guess) are open to misuse. We have a gaping hole between what's possible, technologically, and what should be controlled, morally, to breach yet. And that applies to studies and practices related to genetics, gene therapy and how we treat the outcomes of these experiments: more human than a human? 

And my reference to Skynet, well... I think I shall be watching this purely for entertainment value:

[Anonymous]

I'm not a fan either and what concerns me most is that some security vendors do seem to be tapping AI as a solution to cybersecurity challenges.

I think this is the most likely game over scenario because one AI driven security service will be tasked with protecting another from internal as well as extermal threats as those within an organisation can be as motivated to wreak havoc as those outside.