Media is always full of cases where that AI fails. And in many cases, people expect too much from those tools. AI is not intelligent. It doesn't make new solutions. When the AI sees a program, it selects from the event handlers the solution.
That meant for that of case. Every single case where AI needs to react requires its unique event handler. And the event handler requires a description of the case.
The event handler is the bite of code that determines how the system must react to cases that the system face. If there is no event handler for some cases, that means the system cannot react to that case.
The biggest problem is that every possible case requires its event handler. if the robot must operate independently on the streets. It requires a large number of event handlers.
Driving a car is easier to AI than just walking. Because in traffic are strict regulations for the actions. That means programmers must not create so many event handlers for the car. But that kinds of systems are at prototype levels.
When AI discusses with people it just asks search engines about the things mentioned in the discussions. Then it connects texts from a certain number of home pages. The accuracy of what the AI uses determines what kinds of answers the AI gives.
Older social AI selected answers from databases made for it. And if the user asks the wrong question the AI says: "Sorry I don't know the answer". Or it asks the user to tell the answer to the AI. In that model, the user teaches the software or makes a new record for the AI.
The future is always hard to predict. Trends are changing in the ICT business very often. And the big headlines about AI cause situations in that many corporations invest in that technology. Things like ChatGPT and the Microsoft Bing-cooperative AI are interesting. And social AI brings publicity along with new users for the Bing search engine, and those social AIs are effective tools for marketing.
AI-based software can do many things. They are excellent tools for many things like making automatized programming using AI is the hottest thing in ICT. And if companies don't follow trends and turn boring they will die. That's why trends have a special meaning in the ICT business. And it's a virtual world where people make business by selling code.
In the ICT business like in all other business sectors companies need new products all the time, that they can offer to customers and leads. The AI is not yet ready. It's a product that requires lots of work. But that is the purpose of the R&D work. The ChatGPT and its colleagues are nothing more than algorithms and databases. That is offering information for next-generation AI. But does AI mean the next bubble in the ICT business?
In the past things like virtual game worlds and virtual meeting rooms were hot things in the ICT. Virtual reality and especially net-based virtual reality is also rising. The hybrid conference rooms where a person walks in a game-type environment and goes to a room, where colleagues are introduced as screens on virtual walls are reality.
The reason, why mobile multimedia tools, like Microsoft HoloLens, are not as successful as they should be is that if a person uses those VR glasses while jogging that could cause falls. But otherwise, those HoloLens can connect to an infrared camera, or drones like quadcopters can transmit their images to those screens.
Those systems can also connect with satellites, and people can share their action camera images by using mobile telephone-based applications. But civilians have very little use for that kind of tool. Repairment workers could use those things if they connected with wrist cameras. And that helps them to see things that are behind some corners.
But there are always new trends in ICT. Then those things became old -fashion. Well, all new tools and applications require lots of data. They require the ability to handle very large data masses. The faster internet and more powerful computer technology mean that the systems can handle more and more complicated code. And that allows the creation of new and more powerful AI-based systems.
But then somebody says that AI is the risk. In the same way, we can find risks anywhere. The biggest security risk is other people. So the AI itself is not at risk. The thing that makes risks are people who allow AI to make things that are risking people.
If the AI software is not completed, and it drives the car, that causes a very big risk to the environment. Also, if humans give weapons to the AI it turns dangerous. The situation is similar like we would give a gun to some child. The thing is that AI is not intelligent. It does only things that programmers made to it. This is the thing in the AI and development work.
https://interestandinnovation.blogspot.com/
Comments
Post a Comment