Los recientes avances en IA se suman a la urgencia de la transformación digital en un mundo híbrido. Los educadores se encuentran, a diario, ocupados navegando en la alfabetización en IA, mientras que al interior de sus aulas, se enfrentan a la necesidad de adquirir un conjunto de habilidades de enseñanza y aprendizaje, en torno a la inteligencia artificial y a la escritura con IA.
Como extensión de la alfabetización digital, el concepto de “alfabetización en IA” ha sido definido por Laupichler et al. como: “La capacidad de comprender, usar, monitorear y reflexionar críticamente sobre las aplicaciones de la IA, sin ser necesariamente capaz de desarrollar modelos de IA por sí mismos”. Abarca las capacidades de la IA y la voluntad de que esta herramienta contribuya a informar en el proceso de enseñanza y aprendizaje. En el caso de los educadores, puede consistir en adoptar una tecnología educativa basada en IA, para lograr una calificación más rápida y justa y proporcionar a los estudiantes retroalimentación más sólida y en tiempo real. Para los alumnos, en tanto, puede significar aprovechar las herramientas de escritura con IA como ChatGPT y potenciar el pensamiento de alto nivel.
Para garantizar que educadores y estudiantes puedan utilizar estas herramientas con confianza y responsabilidad, la alfabetización en IA es clave. Vamos a explorar cómo desarrollar la alfabetización en IA en tu institución, sin importar si eres de aquellos que fueron pioneros en adoptar esta tecnología o si eres más cauteloso y recién te estás acercando.
Superar la inacción y adoptar la alfabetización en IA
Naturalmente, el primer paso para aprender a utilizar la IA es siendo un usuario activo. En la actualidad, el sector educativo se encuentra en un terreno un tanto desconocido para las recientes y revolucionarias iteraciones de la IA, en particular, con los (LLM) o gran modelo de lenguaje que componen las herramientas de escritura con IA, como ChatGPT. Aunque los educadores están probando estas herramientas en masa para comprender de mejor manera su impacto, el tener que establecer los parámetros aceptables de una IA, puede resultar desalentador. Hemos visto una gran variedad de respuestas de las instituciones a la nueva generación de escritura con IA, como las reacciones “instintivas”, en forma de prohibiciones y restricciones absolutas, incluyendo una serie de cambios preventivos, con el fin de volver a los exámenes presenciales para disminuir la “amenaza” percibida. Sin embargo, un denominador común de las instituciones ha sido el retraso en la toma de decisiones acerca de lo que constituye un uso aceptable de esta tecnología, así como la consiguiente presión para incluir de forma significativa la IA en los códigos de honor y las políticas de integridad.
Today’s AI-generated text is more disruptive than its normalized predecessor - predictive text input - and is a wake-up call for education to incorporate AI technology into strategy and governance more proactively. It’s clear that students are not prepared to wait, and are already harnessing open-source AI to gain a leg up in their studies. Without direct guidance on technology integral to their future, they risk academic misconduct and poor habits that undermine their goals and learning potential.
Moving from awareness to understanding of AI technology means unpacking it in real time. For instance, educators who have already experimented with ChatGPT to gauge its strengths and weaknesses and what it means for proof of learning in their classroom, are demonstrating the critical evaluation that is instrumental to AI literacy. As a result, they’re in a better position to flip the technology in their favor to support assessment in the new AI era.
Embracing AI literacy enables teaching
Look beyond the open source AI generators dominating the landscape, and you’ll notice a growing precedent of AI and machine learning that improve learning outcomes and enhance student and educator experiences. Embedded in pioneering edtech software, AI’s benefit has typically been strongest in the arena of grading and assessment. For instance, the AI-assisted grading and Answer Groups features in Turnitin’s Gradescope have yielded dramatic time savings during individual and collaborative grading for many instructors and their departments.
However, there is still much work to be done in achieving universal adoption of AI solutions, with institutional preparedness and investment serving as major predictors of AI literacy - particularly in education sectors with centralized buying decisions, and for developing countries with limited resources. Edtech is set to grow by 13.6% from 2023 to 2030 - reaching USD 348.41 billion by 2030 - which is reflective of the need for scalability in teaching methodologies as student numbers rise, and AI’s role in helping educators do more in less time.
What happens if we don’t adapt to AI quickly enough? The World Economic Forum has suggested that without universal AI literacy, AI will ultimately fail us. Aside from making a space for AI in an institution’s strategic vision, making AI technology a part of daily practice means ensuring it is fit-for-purpose across the ecosystem. It also means addressing diversity of experience to encourage uptake and mitigate hesitancy. This includes educators who are not digital natives and may incur additional barriers to acceptance and skill acquisition, plus the new generation of educators who require training to infuse AI principles with existing pedagogy.
How can educators teach AI literacy?
Many educators would agree student assessment is ripe for change, and AI-powered tools signify much more than just efficiency. The next frontier in the quest for ‘smart’ classrooms is using AI to achieve the long-coveted goal of learning personalisation and to better identify at-risk students. We’ve already seen the potency of AI for computer science in delivering real-time, iterative feedback, and there is potential for AI functions and data to assist educators across all disciplines in supercharging their teaching and driving student progress.
Educator-focused initiatives to build AI literacy:
- Testing of ChatGPT and other AI applications to enhance teaching workflows
- Peer learning amongst academics through communities of practice
- Consulting with learning designers who understand AI principles
- Discipline or department-based learning sessions and championing of use cases
- Leveraging educative AI resources; like these Turnitin assets
Understanding AI in students’ learning goals
Students across the world couldn’t believe their luck when ChatGPT hit the scene, and many wasted no time in accessing the open source tool to complete essay assignments in a matter of seconds. As the dust begins to settle, education is at a crucial moment in which to set the tone for students' long-term engagement with AI and determine whether it’s treated as an opportunity to cheat or a tool to empower their learning.
In preliminary survey research on ChatGPT by Study.com, they discovered that 89% of respondents reported having used the tool to help with a homework assignment. Given the temptation and reward it presents, it may come as a surprise that of the 1,000+ college students surveyed, 72% believe that ChatGPT should be banned from their college's network. This data indicates that students already grasp the significance it holds for proof of learning, and it highlights the importance of educators promoting healthy, fair and equitable AI use to empower learners during their academic journey and toward career readiness.
Students are also gleaning AI benefits via edtech tools and the formative learning they support, but this is often applied inconsistently and at the discretion of a trailblazing educator(s) or department, rather than the default at institutions. The rise of LLMs is an opportunity to expand the horizon for AI assistance overall, and for students to engage more meaningfully and co-create the criteria for learning success.
AI writing extends the potential of feedback by positioning students as agents of feedback rather than passive recipients - ultimately putting them in the driver’s seat to steer AI output according to their objectives and become better collaborators. The contextual cues LLMs require from students to prime the auto-generated text, dovetail into lessons on critical thinking, bias, citations, source credibility and originality. Evaluating AI literacy and informing the associated assessment rubrics may involve students priming the response and putting their own spin on the output, or analyzing the AI attempt to tease out higher-order thinking.
Acquiring AI literacy will meet the more immediate need of setting parameters around academic misconduct, but longer-term, will also help pave the way for updated understandings of student originality. LLMs draw from a database of existing content and ideas - and at least for now - yield largely generic responses. Innovative and out-of-the-box thinking will only grow in value, prompting educators to continue to champion these ideals amongst students while balancing the efficiencies of AI authorship.
Shaping the legacy of AI literacy
AI holds immense potential to transform how we engage students in authentic learning and overcome educational shortcomings, but it could run away from us without safeguards in place.
The consequences of passive use of AI versus a proactive approach that champions AI literacy, are markedly different. The latter, which plays to AI’s strengths and weaknesses, is human-centered and offsets risks of educator redundancies, or creative atrophy in the case of writing and ideation.
Acknowledgement of AI assistance - particularly as it grows in sophistication - is important in the convergence of human-generated vs AI-generated output. One solution is Turnitin’s newly-released ChatGPT detection tool, which operates at a 97% success rate and a less than 1% false positive rate. It helps educators perform detection work instantly upon submission, and thereby assess students’ depth of learning. The team is hard at work on AI detection for competitor LLMs, so stay tuned for updates!
There is still much to be determined in the trajectory of AI technology, but staying on top of its development will allow educators to manage its impact on learning. Perhaps this anonymous educator feedback from the aforementioned study.com survey says it best: “As always in education, I think students will still get out what they put in -- those who work harder will learn and achieve more, and those who cheat to get by will have to 'pay the piper,' either sooner or later.”