Artificial Intelligence 2020 - is there progress?

 

I was very impressed by this article on hubrahttps://habr.com/ru/company/smileexpo/blog/409051/ from 2017. It's been 3 years and what's the progress? Where is the conversational AI? Why hasn't it been seen or heard yet? Such technology could hardly be classified or hidden. Have they not moved on?

When I looked through the article, it was as if I was reading my own thoughts expounded by someone else. Only there is not a penny of philosophy, but the level of understanding of everything else is off the charts. That's why I wonder why there is nothing complete yet... Isn't there enough philosophy?)

Дмитрий Муромцев (ИТМО) — об онтологическом моделировании и формировании разговорного интеллекта
Дмитрий Муромцев (ИТМО) — об онтологическом моделировании и формировании разговорного интеллекта
  • habr.com
Дмитрий Муромцев, руководитель международной лаборатории «Интеллектуальные методы обработки информации и семантические технологии» ИТМО и заведующий кафедрой ИПМ, рассказал о сущности онтологического моделирования, об использовании графов знаний в бизнес-процессах и о работе по созданию разговорного интеллекта. Интервьюер: Анна Ангелова (А.А...
 
Does Yandex's Alice count?
 
Evgeniy Zhdan:
Does Yandex's Alice count?

Well, hardly... There is a much bigger plot in the article. Alice suffers from a short memory and doesn't keep the conversation in the context of the topic or the history of the whole dialogue. She clings to the standard chatbot format of request-answer. Any non-standard queries are forwarded to a search engine. Its engine has no ties to the physical world. In short, it has no hint of knowledge graph technology. That's what it seemed to me when dealing with it.

 
If we are talking about conversational AI, they are implemented quite well in modern promotional bots, including those for shops. Models close to release and installation in shops recognise faces, remember visitors, remember facts of the conversation (and purchases), and talk accordingly. The biggest problem so far is when the promobot is surrounded by several people and trying to talk at the same time. I've seen stuttering when doing this.
 
I was teased hard a few years ago at a children's exhibition. A girl so expertly controlled a robot and imitated communication that I believed in artificial intelligence. I was talking to a robot, an intelligent robot. I couldn't believe my eyes. Googling it later at home, I experienced strong emotions. How could a grown man be so naive


 
Edgar Akhmadeev:
If we are talking about conversational AI, they are implemented quite well in modern promotional bots, including those for shops. Models close to release and installation in shops recognise faces, remember visitors, remember facts of the conversation (and purchases), and talk accordingly. The biggest problem so far is when the promobot is surrounded by several people and trying to talk at the same time. I've seen stuttering when doing this.

I wish I'd had the chance to talk to such "promobots").

 
Aleksei Stepanenko:
I was brutally scammed a few years ago at a children's show. A girl so skillfully controlled a robot and imitated communication that I believed in artificial intelligence. I was talking to a robot, an intelligent robot. I couldn't believe my eyes. Googling it later at home, I experienced strong emotions. How could a grown man be so naive


The easiest and most effective method of identifying a true AI is to ask questions that ask for explanations of the physical causes of ordinary everyday phenomena, for example: "explain the physical reason why cows can't fly?". A human can easily find the answer, but no chatbot can give a correct and adequate explanation.

 
Реter Konow:

Why can't cows fly?

Afraid of heights, eh?

 
Aleksei Stepanenko:

Scared of heights, eh?

Now I know for sure you're not a bot! ))

By the way, it's more psychological, and probably even psychiatric in nature. And the physical one is a bit on the other plane. So perhaps you are not a bot, but an alien.))
 
Реter Konow:

For example: "explain the physical reason why cows can't fly?".

If AI is not based on linguistics, but on semantic forms, it can explain more than that. And in principle, this question has long been solved. There are many things that are not advertised and even if it is brought to public discussion, it will not do any good. No one will simply fit it into the usual logic.

As far as I am concerned, the development of AI has not done humanity any good. And wrong are those who think it's interesting and exciting. It would be better if these technologies had not progressed.

 
Реter Konow:

...For example: "explain the physical reason why cows can't fly?". A human can easily find the answer, but no chatbot can give a correct and adequate explanation.

By the way, the "crafty" Alice would answer something like "what do you think?"