The VRAI environment is the virtual learning space made possible by the imminent maturation and convergence of virtual reality and artificial intelligence. We foresee this as a possibility within ten to fifteen years. The VRAI environment is not a game, although it incorporates some aspects of gaming. It is a virtual reality environment, like a much more sophisticated version of Second Life. VRAI will be offered in all common languages.
It's 2030.
Hardware
Although the VRAI environment can be viewed on a computer monitor, in order to experience the VR aspect, users generally use a VR headset. By now, these headsets are light weight, wireless and contain sufficient computing power so that they can be interfaced directly with the Internet instead of being tethered to computer.
Visual Environment
Processing speeds and graphic display capabilities are such that images can be rendered in ultra-high definition and are virtually indistinguishable from the real world. Of course, the VR headset provides a true sense of depth.
There are wide range of visual settings available within VRAI, from urban streets and buildings to rural scenes. Interior spaces offer the same possibilities that can be found in the real world, for example, offices, restaurants, stores, family homes, transportation, etc. Due to its self-generating graphics engine , VRAI can create visual environments without additional software programming . As the goal of the project is to contextualize language learning, the particular visual
setting that the user finds herself in is contingent upon the task that the learner pursuing . These visual settings are populated both by AI-controlled characters and user-controlled characters.
Role of AI
Language learning requires interaction, ideally, with native speakers of the target language. In traditional classrooms, one teacher provided language modeling for an entire class and when students practiced speaking the language, it was with each other. This scenario provided students with no correct language modeling during conversation.
At the heart of VRAI is its AI engine. With the exponential rise in the power of AI in the 2020's came the ability for human to computer interaction in ways that rival human to human interaction. By this point, computers are adept at natural speech processing. They are able decipher and understand human speech as well (and sometimes better) than most humans and are able to produce speech without any of the synthetic inflections that was so characteristic of the first couple of decades of the century. Through years of training the AI, it is able to produce speech that sounds just like a native speaker and can express itself using a variety of authentic regional accents.
The problem of providing correct language modeling has been solved. Learners interact with AI controlled characters within VRAI in the same way the do with human users.
The implementation of AI goes far beyond understanding and producing speech. The AI is responsible for guiding the learner through the entire language learning process.
Navigating the Environment
Recognizing that language learners may be starting with absolutely no prior exposure or competency in the target language, there needs to be a significant induction stage throughout which the learner will acquire the basic vocabulary needed for common, simple interactions such as introducing oneself, greeting, asking politely for an object, excusing oneself, etc. These simple interactions will need to be mastered before the moving on to engaging in meaningful conversation. In these entry-level activities, the learner will have an AI controlled virtual guide to scaffold the learning. For instance, to practice introducing oneself, the guide accompanies the learner in a situation where they meet a group of people (AI controlled). The learner watches and listens to the guide as he introduces himself to several people and then the learner takes her turn, attempting to use the same language that was modeled.
The AI is constantly attending to the vocabulary, pronunciation and grammatical structures used by the learner. When errors or inaccuracies occur, the program provides additional positive language modeling and creates more opportunities for practice. Through continuous formative assessment, VRAI knows when to begin to push the learner to more challenging language situation. Once the learner has acquired some basic communication skills, VRAI is able to place the learner in more complex communicative situations.
Learner Profiles
As mentioned, VRAI is constantly tracking learner interactions, looking for strengths and weaknesses. This allows the software to create a complex profile of the learner's progress. In addition to the profile informing the choices the the AI makes in order to customize the user's experience to their individual needs, the learner can also consult the profile which provides a summary of progress in specific language areas. For example, within the profile the user can choose the category comparing and contrasting, and then view a bar graph indicator of progress in terms of vocabulary and structures. Choosing details bring up data on specific vocabulary understood and produced, structures correctly used and errors that were made.
Modes
Users will have the option of using VRAI in adventure mode or task mode. Adventure mode allows learners to freely explore the environment, interacting with characters and objects of their choosing. Of course, the AI has already developed a sophisticated language learner profile of the the user and consequently, users are presented with situations that they are able to navigate, but which contain some degree of challenge.
When users choose task mode, they will be able to choose from a variety of language skills to practice. Because the AI has created a profile for the learner, it is able to suggest level appropriate tasks. Unlike certain games, VRAI will not lock content until prior content is successfully completed, it will, however, advise the learner that he may find it quite difficult if he chooses a task that is much beyond his linguistic competencies as reflected in the learner profile.
Assessment
Summative assessment is unnecessary in VRAI thanks to the sophisticated ongoing formative assessment. Progress is logged in the learner profile which provides easy access to learning metrics.
It's 2030.
Hardware
Although the VRAI environment can be viewed on a computer monitor, in order to experience the VR aspect, users generally use a VR headset. By now, these headsets are light weight, wireless and contain sufficient computing power so that they can be interfaced directly with the Internet instead of being tethered to computer.
Visual Environment
Processing speeds and graphic display capabilities are such that images can be rendered in ultra-high definition and are virtually indistinguishable from the real world. Of course, the VR headset provides a true sense of depth.
There are wide range of visual settings available within VRAI, from urban streets and buildings to rural scenes. Interior spaces offer the same possibilities that can be found in the real world, for example, offices, restaurants, stores, family homes, transportation, etc. Due to its self-generating graphics engine , VRAI can create visual environments without additional software programming . As the goal of the project is to contextualize language learning, the particular visual
setting that the user finds herself in is contingent upon the task that the learner pursuing . These visual settings are populated both by AI-controlled characters and user-controlled characters.
Role of AI
Language learning requires interaction, ideally, with native speakers of the target language. In traditional classrooms, one teacher provided language modeling for an entire class and when students practiced speaking the language, it was with each other. This scenario provided students with no correct language modeling during conversation.
At the heart of VRAI is its AI engine. With the exponential rise in the power of AI in the 2020's came the ability for human to computer interaction in ways that rival human to human interaction. By this point, computers are adept at natural speech processing. They are able decipher and understand human speech as well (and sometimes better) than most humans and are able to produce speech without any of the synthetic inflections that was so characteristic of the first couple of decades of the century. Through years of training the AI, it is able to produce speech that sounds just like a native speaker and can express itself using a variety of authentic regional accents.
The problem of providing correct language modeling has been solved. Learners interact with AI controlled characters within VRAI in the same way the do with human users.
The implementation of AI goes far beyond understanding and producing speech. The AI is responsible for guiding the learner through the entire language learning process.
Navigating the Environment
Recognizing that language learners may be starting with absolutely no prior exposure or competency in the target language, there needs to be a significant induction stage throughout which the learner will acquire the basic vocabulary needed for common, simple interactions such as introducing oneself, greeting, asking politely for an object, excusing oneself, etc. These simple interactions will need to be mastered before the moving on to engaging in meaningful conversation. In these entry-level activities, the learner will have an AI controlled virtual guide to scaffold the learning. For instance, to practice introducing oneself, the guide accompanies the learner in a situation where they meet a group of people (AI controlled). The learner watches and listens to the guide as he introduces himself to several people and then the learner takes her turn, attempting to use the same language that was modeled.
The AI is constantly attending to the vocabulary, pronunciation and grammatical structures used by the learner. When errors or inaccuracies occur, the program provides additional positive language modeling and creates more opportunities for practice. Through continuous formative assessment, VRAI knows when to begin to push the learner to more challenging language situation. Once the learner has acquired some basic communication skills, VRAI is able to place the learner in more complex communicative situations.
Learner Profiles
As mentioned, VRAI is constantly tracking learner interactions, looking for strengths and weaknesses. This allows the software to create a complex profile of the learner's progress. In addition to the profile informing the choices the the AI makes in order to customize the user's experience to their individual needs, the learner can also consult the profile which provides a summary of progress in specific language areas. For example, within the profile the user can choose the category comparing and contrasting, and then view a bar graph indicator of progress in terms of vocabulary and structures. Choosing details bring up data on specific vocabulary understood and produced, structures correctly used and errors that were made.
Modes
Users will have the option of using VRAI in adventure mode or task mode. Adventure mode allows learners to freely explore the environment, interacting with characters and objects of their choosing. Of course, the AI has already developed a sophisticated language learner profile of the the user and consequently, users are presented with situations that they are able to navigate, but which contain some degree of challenge.
When users choose task mode, they will be able to choose from a variety of language skills to practice. Because the AI has created a profile for the learner, it is able to suggest level appropriate tasks. Unlike certain games, VRAI will not lock content until prior content is successfully completed, it will, however, advise the learner that he may find it quite difficult if he chooses a task that is much beyond his linguistic competencies as reflected in the learner profile.
Assessment
Summative assessment is unnecessary in VRAI thanks to the sophisticated ongoing formative assessment. Progress is logged in the learner profile which provides easy access to learning metrics.