Samsung recently introduced Bixby, it's attempt to create a praise worthy AI assistant. Previous attempts have failed, and therefore a lot is expected from this system, although Samsung placed it in the sphere of intuitive interface, not personal assistants. Although it's been created by associate company Viv (which has, prior to the Samsung acquisition, created the Apple Siri), Bixby will not be able to cope with Siri, Alexa and other AI systems. But, that doesn't matter.
The AI assistants can also be found on the websites and in telephony, in the form of chatbots or automated voice messages. The latter form is known as the "answering machine", and it's been present for around twenty years. But again, it is worth wondering how in such a short time virtual assistants evolved from recorded messages, which are activated by pressing the designated button, to omniscients who are ready to do almost anything we ask at any moment.
Machine learning and artificial intelligence
The most important things here are essential two concepts that correlate with each other and so create "personality" of virtual assistants. One is machine learning, or the ability of computers (and other software devices) to learn or acquire data without first being programmed. Artificial intelligence is the one that the machines create, or generate the perception of the space around them with the execution of tasks, which seeks to successfully achieve the goal. The combination of these two possibilities are obtained by virtual assistants who are able to carry out the task within the capabilities of the device on which they are located, and have a broad base of terms which they handle.
For example, on the phone, you can activate the assistant by voice command or by clicking, that will tell you what the weather will be tomorrow, give advice what to cook for dinner to your partner, turn on your favorite music, or tell where to find motorcycle parts. On the smart home systems, they, by a similar principle, regulate the heating, lighting, hot water, and you can tell them when you'll get back home, and give them a command that they shall perform. All responses that are generated by AI assistants are determined by individual habits. For example, if you are looking for pasta on sale, the assistant's first choice for food is offering pasta or a good Italian restaurant nearby. Of course, the assistant is connected to other systems in the device, ie. geotagging, so it will recommend a place where you already have been, especially if you gave it a good rating.
Do we need them?
As we know everything they are capable of, it's worth asking to what extent AI assistants are useful and whether we need them. In any case, they may be useful in a time when we are engaged in another activity, when we are undecided, or when we don't know what to do. But that does not mean that we can completely rely on them.
As we said, virtual assistants absorb the events from their nearest environment, and given the fact that they come without any knowledge (except for those who have direct access to services such as Google), it is difficult to consider that AIA's will know everything in a given environment. Here's also the question of morality, because if the machine makes a mistake, it will not feel guilt or any other emotion, but will continue to work as if nothing had happened, believing that it's is the best solution.
For example, Siri has a problem with the recognition of the words "e-sports" and "escort" and by mentioning of one term, you will often get an answer to the other one. Microsoft's chatbot Tay, surrounded by various users from polite and communicative character turned into a sexist, racist, Nazi and almost everything else that society considers negative. Sometimes, you need to shout at Alexa, because it will not stop the activity for which is clear that it's not appropriate to a given task. Sometimes, except bugs, it can be said that AI assistants act just like people.
Due to the fact that people are not perfect, nor do they know everything, then no virtual assistants will know everything, nor be perfect. It would be unrealistic to expect such a thing, but their usefulness in certain moments, when the human brain has no adequate solution for a situation, is very welcome, but only to a certain extent. In fact, people should not allow the abandonment of cognitive and thinking tasks, and let those to machines, because it, in theory, could lead to errors, or reduce human intellectual potential, or, in other words, people will become unresourceful, they wont think, and generally have reduced logical, reasoning and understanding skills.
More about AI.