The use of artificial intelligence is increasingly spreading beyond a narrow circle of tech enthusiasts. Now this is a mass product that companies are actively incorporating into their services, especially virtual assistants.
Google and Microsoft have already integrated generative AI into their assistants. Amazon also plans to launch an AI-based assistant this summer.
However, even standard virtual assistants have a very wide range of functions – they can play a fairy tale, find information on the Internet and even close the curtains or boil the kettle.
first assistants
Modern virtual assistants appeared more than 10 years ago. The starting point is Apple’s presented Siri. This assistant can give you weather information, play songs, and even remind you to turn off the stove via voice command. This was a real revolution; Before this, no device could recognize tasks at such an advanced level.
Immediately after Apple, other technology companies also introduced their virtual assistants. on Amazon in 2014 showed a “talking” Echo speaker with a built-in Alexa assistant – it had about 50 thousand skills, the most important of which was controlling smart home devices. In particular, she could adjust the lighting in the house, order groceries for home delivery, and solve other simple household tasks. The device turned out to be as easy to use as possible, so even a non-advanced user could immediately feel the comfort of a smart home. The speaker quickly became popular – in just two years the company managed to sell more than 5 million devices. After this, Google decided to do the same and introduced the Google Home speaker in 2016. The success was surprising; Since its launch, the company has been selling one Google Home Mini speaker every second.
Although Russian companies began to conquer technology later, they took it no less actively. For example, Yandex introduced Alice, which can communicate, in 2017. A little later, a virtual assistant appeared in VK. “Marusya” can collect shopping lists, check the strength of passwords, play with children and offer interesting activities when you are bored. Sber, on the other hand, has launched a whole family of virtual assistants, each with a unique personality, voice, and communication style. In this way, the user can choose an assistant that is close to his/her soul.
Answers according to scenarios
The first modern virtual assistants, such as Siri and Alexa, worked based on built-in action scripts. The user asked, “Will there be a snowstorm today?” When asked a question like , the assistant first recognized the speech and then special software converted the voice command into text format and sent a request for processing. The algorithms then determined the subject of the request (in this case, weather) and searched the internet for the necessary information. Then the answer given by the assistant was: “A snowstorm is expected and it feels like minus 15 to 10 degrees below zero.”
Virtual assistants have evolved significantly over the last decade. For example, they acquired an advanced natural language processing and recognition engine that allows them to clearly understand the context of the request. For example, you can ask to be reminded of an appointment in several ways: “remind me of the meeting at 11 am tomorrow” or “notify me of the meeting at 11 am tomorrow”, but the assistant will correctly interpret your request and fulfill it. BT.
Until recently, assistants were used only to perform simple search queries and simple daily tasks. Conducting a meaningful dialogue and expressing their own views remained an impossible task for them.
While the capabilities of virtual assistants have expanded with the number of applications and devices they can interact with, their actual capabilities have remained limited. For example, by turning on a flashlight or camera, the assistant performs the same task and only interacts with different applications. Yes, the ability to play music or set a voice reminder pleased users, but such monotonous functionality could no longer surprise anyone.
GenAI answers
In 2022, Open AI’s Chat GPT became a sensation, overnight forcing all major technology companies to refocus on the development of productive AI technologies. This is a type of artificial intelligence that can create new content based on the data it has been trained on. Realizing the potential of the technology, companies have begun to implement generative AI in virtual assistants on a large scale. First of all, this technology has opened up the possibility of taking assistants beyond simple voice requests. Second, I added the ability to learn. This means that the virtual assistant will now be able to adapt to the needs of specific users.
Nowadays, everyone from small startups to large technology companies is specializing in generative AI. Meanwhile, this time Russian companies realized the potential of the technology faster than their Western counterparts. Thus, it became the first company in the world. Done Generative AI in its voice assistant has evolved into Yandex. In May 2023, “Alice” received an update in which she learned to find text, generate ideas, and solve other complex problems. And Sber in November translated “Hi” to your virtual assistant GigaChat neural network model. It is trained on a large corpus of Russian texts, which allows it to better take into account the nuances of our language and the specific needs of Russian-speaking users.
Silicon Valley leaders are just starting to incorporate generative AI into their assistants. Google in the fall of 2023 Done transfer this technology to your voice assistant and Apple statedSiri will get productive artificial intelligence this summer.
It’s not that simple
Ever since rumors about generative AI took the world by storm, major tech companies have poured all their resources into becoming leaders in this field. At the same time, the industry’s appetite is so high that many companies are ready to sacrifice technology and employees in other departments just to find additional funds. In January 2024 alone, approximately 23,670 people were laid off from 85 technology companies; this is the highest figure since March 2023. Even Apple hasn’t been able to seamlessly refocus on developing services with built-in AI. Less than a month ago, the company closed the project to create driverless cars and transferred part of the team working on it to productive artificial intelligence development projects.
There is only one reason for this: creating a great productive model requires huge investments. Applying technology to a virtual assistant also comes with a cost. At the same time, the short-term return on all these investments is questionable. The user who purchases a smart speaker can use it for 5 years or more and does not pay anything to the virtual assistant. As a result, it turns out that the company spends money on the development and improvement of virtual assistants, but makes a profit only from the sale of devices and content services on which they are already installed.
Additionally, with the emergence of generative artificial intelligence, assistants are becoming much smarter. Their answers are no longer tied to scripts and templates: They can endlessly generate ideas and perform other tasks previously only possible for humans. That’s why companies decided to look for ways to make money from advanced assistants.
That’s why in the fall, Amazon announced that a new version of the Alexa Plus voice assistant with built-in generative artificial intelligence will appear in the summer of 2024 and will be available only with paid subscription. Raising additional funds should cover the company’s expenses for developing and improving the user interface. Microsoft follows a similar strategy; The Copilot Pro virtual assistant is currently only available to users as part of a paid subscription.
Just a few weeks ago, Google launched its AI-based Gemini assistant. opened Gemini Ultra subscription, an even more advanced assistant model.
Wait and see
The productive artificial intelligence market is now It is evaluated The future of $11.3 billion virtual assistants looks promising. Experts predict that they will soon be able to better understand context, read user emotions, and even change their tone of voice depending on the topic of the conversation. It’s too early to say that standard virtual assistants without built-in productive AI will disappear completely. However, it is quite possible that companies will become more active in introducing paid subscriptions to assistants with “superpowers”. Users will soon be able to obtain assistants whose abilities will be as close to human intelligence as possible.