AI and Education
Designing AI to Enhance Rather Than Undermine Learning:
One profound risk to education is students' use of AI to execute written tasks rather than learning, synthesizing and analyzing information and ideas themselves and developing the learning, organizational, and writing skills to complete tasks themselves.
This will hinder student development in terms of their ability to learn, to think, and to work and serves no obvious benefit aside from increasing the efficiency of output and the return of satisfying grades, which would in this case not be an indicator of knowledge or ability.
◇ Teachers can and should be used to help design AI systems to serve educational purposes in order to ensure that the systems are designed in such a way to support rather than undermining student learning and effort.
A strategy in designing educational AI is to build a system which requires frequent and significant human feedback, by putting humans in the loop this ensures that actual learning and skill development is still required of students.
◇ Teachers can and should be involved in decisions about what material is used to train educational AI and how the systems are operated.
AI as a Learning Tool:
◇ AI could be helpful in education as a tool for multilingual individuals, enabling them to translate information from their language of greatest ease into the language in which they are working or studying thus helping them to be more efficient learners and communicators.
The caveat is that this would hinder the development of their language skills and fluency in their secondary language, rendering them dependent on translation tools rather than on personal knowledge and ability.
◇ AI can be a tool for students to pursue questions and areas of confusion in greater depth than a teacher managing many students may have time or, and may be able to provide students with a form of information which suits their specific needs at that time.
◇ Can be used to efficiently suggest and provide resources for further information and study, in this case functioning as a mechanized librarian or archivist.
◇ Creating AI that is suited to context, meaning the needs of a specific group of students based on culture, location, learning and language challenges and more.
Testing in schools is a way for teachers and administrators to gauge the progress, shortcomings and needs of students in the course of their education. It serves as an indicator of student learning, the effectiveness of educators, and the quality of the school district. Test scores are also used as an indicator of students ability to perform at certain levels, and to project future abilities: SAT and ACT tests are used as indicators of future college performance.
However, test scores most fundamentally represent opportunity rather than ability, meaning the access that a student has had to quality education throughout their learning career.
Stealth Assessments: This is one technique for educators to evaluate students actual knowledge and ability, by surprising students with tests in order to prevent the use of AI systems to improve student performance. Stealth assessments can also take the form of small assessments embedded in lessons and ordinary class activities, which the students do not know are assessments, this again prevents students from skewing their performance by the use of technological enhancements.
Scoring and Evaluation: AI systems are used to process, evaluate, and score student assessments rapidly. This is used in terms of statewide evaluations and for online courses. AI systems are used to evaluate not only bubble chart tests but also the written and essay sections of examinations.
◇ One advantage to this is that it decreases the costs of assessments and potentially also of courses because there is less time required of educators to score and evaluate the tests.
The caveat is that this also means less employment for educators who would otherwise be involved in evaluating assessments.
Generating Assessments: AI systems are also being used to generate assessments altogether, producing variations on questions to prevent redundancy and student cheating.
◇ AI could potentially be used to create questions and assessments based on themes of interest for students, potentially making it easier for the student to understand the questions.
Automatic grading: The AI system Bakpax envisions a system in which teachers can take a picture of student homework which will then be automatically graded by vision based AI systems, thus eliminating the time and effort a teacher would normally invest in personally observing a students learning progress,
◇ The caveat is that teachers would then be less familiar with individual student progress, and with the personality and mannerisms of the student which are embedded in how they learn and complete assignments, which may hinder a teachers knowledge of and ability to best teach their students.
Changing Assessments: AI systems are designed in such a way that they perform extremely well on logic and speed based assessments such as the SAT and ACT which are the essential college entrance exams. What this means is that these scores indicate students ability to perform in ways which can be accounted for by AI, meaning that these evaluations are not as relevant or effective as indicators of student ability in an AI world.
Testing will need to be adapted to test for abilities and potential which AI cannot take one, in order to indicate students aptitude in a changing world in which different skill sets will be prized.
The search engine was a tool which was introduced in the 90s and transformed the way in which we seek, access and engage with information. Similar to AI, the introduction of the search engine was a major shift in our relationship to knowledge, information and learning, and it has now become commonplace, and deeply embedded in the ways in which we think and interact with the world.
ChatGPT is a next level search engine, which provides articulated answers to distinct questions, as opposed to a search engine which provides a list of relevant information correlated with the initial question or keywords. This change means that the user cedes decision making about what is important and relevant to the AI system rather than making those decisions themself by browsing the list of possible answers.
This is obviously highly problematic because AI cognition lacks emotional, social, ethical or cultural awareness or standards and so cannot contextualize or discern information based on these qualities.
AI as tutor and Question Answering Assistants: AI can potentially be used to assist teachers in response to student inquiries in needs outside of class, and to provide them support and responses specifically suited to their needs or providing in depth responses to specific interests and questions which are beyond the teachers scope of either knowledge or time.
AI taking on Mundane Aspects of Teaching: One thing this could enable is for educators to concentrate on the culturally, socially, and emotionally specific needs of the class and students while delegating routine tasks to AI systems.
Continuing to Teach Basic Operations that Are Otherwise Automated: Logical operations such as factoring and algebra are of course easily delegated to calculators and calculator applications, however if students do not learn these operation themselves to some degree, then these operations will effectively be relegated the the realm of ‘magic’ and AI will be entirely responsible for and empowered by this ability. This is one of many dangerous instances in which humans are at risk of ceding intellectual ability, understanding and thus agency to machines.
It is essential to remember that AI is predictive and logical, and so is capable of replicating existing content, but this does not mean that it possesses understanding. Complex and contextual understanding is an imperative human ability which is entirely distinct from the cognitive abilities of machines.
AI Compared to the Calculator:
The introduction of the calculator into the classroom and commonplace learning was a major change in the teaching, learning and exercise of mathematics, and can in some ways be compared to the introduction of AI into education and learning. The introduction of the calculator meant that students could engage in solving complex computational problems more quickly and efficiently than before, and meant that although they needed to have knowledge and skill in many mathematical operations, they did not need the same level of proficiency as before.
Chat AI, meaning those natural language AI systems capable of producing written content such as essays, has the capacity to undermine student writing. This is highly problematic as writing is the mode of both thinking and writing and is an essential tool in learning to think, to compose ideas, and to communicate. Chat AI systems may mean that students no longer need to be fully proficient in reading or writing.
However, comparison to the calculator may not effectively indicate the magnitude of AI, which may instead be more comparable to the printing press in the level of change it will create.
Reevaluating and Reshaping of Education:
The advent of AI necessitates a reevaluation of the skills, values, knowledge, and abilities that education should develop in students and the ways in which these are taught. Students will enter a more unstable world in which it will be essential to have the ability to adapt and react quickly and to be resilient and strategic in the face of change. This is a result of both the pace and nature of technological development, and fraught economic and environmental conditions which will only increase in coming years.
Education should foster agency, awareness, critical thinking, problem solving, connectedness, and well-being, all human capacities which cannot be ceded to AI.
◇ A major concern is how education in a world of AI can promote deeper rather than shallower learning.
◇ AI will bring about a major re-emphasis on advanced social and emotional capacities.
◇ Teachers will bear a major responsibility in educating students in ethical principles and critical thinking abilities to ensure that they are agile, thoughtful, moral and self-possessed users of technology, and are capable of understanding and making meaningful and purposeful decisions in a world of excessive automation.
◇ Teachers will also be tasked with making students keenly aware of AI’s profoundly dangerous capacity to reduce human connection and increase loneliness, to ensure that students are aware of this danger which is embedded in their relationship to technology.
A 2020 McKinsey assessment states that teachers currently spend around 50 hours a week working, of which only 50% is spent in direct interaction with students. The delegation of mundane tasks to AI has the potential to increase the time spent focusing on and responding directly to students.
Digital Literacy Education: This will be essential in ensuring that students around the world are able to develop the skills and fluency they need to function in the world of AI. Currently, digital education is neither equitable or widely accessible and this will need to change.
The advent of AI and its relationship to learning and cognition demands a reevaluation of the goals of education and the purpose it serves in adolescent development and workforce preparation.
◇ One role of education is to establish a child’s social and future professional networks and begin to establish their social capital, which are two of the most essential elements in securing a well paying job and higher quality of life.
◇ Recent studies have indicated that relationships across socio-economic lines have greater impact on life outcomes, particularly for professional prospects and income than do test scores or grades.
◇ A potential application of AI for the purpose of improving cross-socioeconomic relationships is in redesigning school districts to mitigate segregation.
Educational Changes During the Industrial Revolution Compared to Changes Resulting from AI:
◇ During the industrial revolution schooling was adjusted to prepare students/children to work in factories, and so art was de-emphasized in curriculum, and this continued for more than a century
◇ The advent of AI may result in a de-emphasis on academic ability and an increased emphasis on specifically human social and emotional capacity which cannot be performed by AI, in schooling and the workplace.
◇ Schools will need to focus on develop complex cognitive and socio-emotional skills, developing creative in such a way as to work collaboratively with AI will be essential, meaning that creative development should be a prominent aspect of education
We all currently use AI in the form of grammatical and syntactical correction, sentence auto-complete, route suggestion on maps, just to name a few.
Continue Reading about the AI transition and the ways it will effect how we live, work, and think.