In my opinion, the most important thing is to first learn about the reality of AI, what it can contribute and always use it with a pedagogical purpose that demonstrates that its use is useful. If we don't know if the activity will really be better with AI than a traditional activity, it is worth continuing with the traditional way until we can find a clear and realistic example that really affects the learning process of the student. Therefore, I believe that AI should be used in an informed way and not simply because it is fashionable.

I liked the idea from one of the webinars I've recently visited about asking students to hand in their essays together with the drafts they created. A way to prove that they put in some effort, anyway.

Another interesting idea from the teacher who uses AI in the writing class to develop students' critical thinking. Students have a topic to write about. They ask AI to do it, but then they have to analyze AI's job. There are questions prepared by the teacher, like: Does the essay have a hook? What is the paper's thesis (highlight in the text)? Highlight the personal examples in the text generated by AI. Do they support the thesis, etc.? So, students not only copy-paste but analyze the text, find mistakes.

Teachers are very creative and the'll find the way out:)

Assessment should also take place through conversation.

For summative assessment, I still prefer traditonal tests.

• Use a variety of assessment methods, including written assignments, oral presentations, group projects, and practical demonstrations.

• Collect and Analyze Data

• Involve Students in Assessment Design

It was obvious, therefore, that biases would end up affecting AIs. There should be a global consensus to tackle these biases and prosecute any AI that is proven to have an aggressive and conscious bias.

As how this affects our task as teachers, I think it's yet to be seen. We have to be careful with how we use AIs and keep a critical eye on everything we use them for.

We should conduct our own individual research to check on the validity, reliability and fairness of the AI-based evaluation tools we choose to use or design.

Maybe, elaborating a correlation of results between what we think the students should receive (both in formative and summative assessment) and what AI provides.

Any thoughts?

In order to adapt our assessment practices and, at the same time, ensure fairness and effectiveness, I think we should use formative and continuous assessment methods that can monitor students' progress and performance throughout the learning process, not just at the end. These methods can include activities such as portfolios, reflections, self-assessments, feedback, among others, which can encourage students to develop their skills.

One way to take advantage of formative assessment strategies to address concerns about the use of AI-generated essays and assignments and algorithmic bias is to promote students' critical reflection on the processes and products of their learning. Critical reflection allows students to analyze their own strengths and weaknesses, their learning strategies, their goals and their outcomes. Critical reflection also helps students to identify and question the possible biases and prejudices that may be present in the algorithms they use or that are used to assess them.

As for the use of AI, we could assign them to explore some subject independently of AI and compare it to what AI has to say - make a comparison; discuss it in school, and then maybe go back to AI and teach it if it was wrong.

Only what we can do is to change approach.

To give students open ended questions, collaborative projects, create assesments which ask chritical thinking, creativity. Learn them how to use AI tools in a proper and critical way.

Looking back to when I was doing my undergraduate research, I used to spend plenty of time in the library doing research in various printed and online publications.  At that time, I also had to read these findings and interpret them with a critical eye so that I do my conclusions and recommendations.

Having said this, AI is facilitating my research and the time I spend looking for relevant literature in the given subject area.  However, I still need to have the skill of critically analysing the generated result.

So, as educators we need to emphasize and train our students to foster critical thinking and ethical awareness of world reality.

Looking at things top- down, namely, focusing on developing the future skills are in demand for the next 5 years, then we will go hand in hand with what machines and algorithms involve.

According to the World Economic Forum, the top 10 required skills of 2025 are: analytical thinking and innovation, active learning and learning strategies, complex problem solving, critical thinking and analysis, creativity, originality and initiative, technology use.

Looking at these skills, it is easier to project the assessment tasks- they will consist on group work , on peer analysis, on grading originality, the project based activities are wonderful in developing the required skills as they involve- initiative, creativity, organizing, technology use, self management , presentation and critical thinking.

It stands in our power to be resourceful, innovative and help our students align with the present requirements.

Remembering that the use of assessment strategies must be adapted according to the age group, educational level and specific characteristics of the students.

It is important that it is explained to students that artificial intelligence platforms are a tool and should not be used entirely to carry out work. It's important to show that are important use books too, going to the library, for example, should be encouraged.

Sources of Bias in AI Assessment

AI algorithms can be biased due to the data they are trained on. If the training data is not representative of the population being assessed, the AI algorithm may learn to make biased decisions. For example, an AI algorithm trained on a dataset of mostly white, middle-class students may be more likely to give lower scores to students from different racial or socioeconomic backgrounds.

Additionally, AI algorithms can be biased by the way they are designed. For example, an AI algorithm that is designed to assess writing may penalize students who use informal language or slang, which can disadvantage students from marginalized communities.

Addressing Bias in AI Assessment

There are several strategies that can be used to address bias in AI assessment. One strategy is to use data augmentation techniques to make the training data more representative of the population being assessed. This can involve techniques such as oversampling, undersampling, and synthetic data generation.

Another strategy is to use bias detection and correction techniques. These techniques can be used to identify and correct for biased decisions made by AI algorithms. For example, an algorithm could be trained to identify and remove features that are known to be correlated with bias, such as a student's race or socioeconomic status.

It is also important to develop AI assessment tools that are transparent and explainable. This will allow users to understand how the AI algorithm is making decisions and to identify and address potential biases.

Promoting Fairness and Equity in AI Assessment

In addition to addressing bias, it is also important to promote fairness and equity in AI assessment. This means ensuring that all learners have access to high-quality AI-powered assessments and that the assessments are used in a way that is fair and non-discriminatory.

One way to promote fairness and equity is to use AI assessment tools to provide students with customized learning experiences. For example, an AI algorithm could be used to identify a student's strengths and weaknesses and then provide them with personalized feedback and recommendations.

Another way to promote fairness and equity is to use AI assessment tools to identify and address systemic inequalities in education. For example, AI algorithms could be used to identify schools that are not providing all students with equal access to high-quality instruction.

As an EFL (English as a Foreign Language) teacher, I think the greatest risk primarily concerns the evaluation of writing tasks, such as essays, articles, e-mails etc. Standard grammar, vocabulary, and language function assessments should not pose a real problem.

Diversifying assessment methods to include a range of tasks that assess different skills and competencies is certainly a good way to reduce the impact of biases and keep up with the times. In the same way, combining AI tools as aids in the assessment process with a human element in the evaluation is fundamental to ensure a more comprehensive assessment.

As for writing assessments, the only way I can think of in order to reduce plagiarism or superficiality is on-site tasks. This means teachers should dedicate some classes to get their students to write a piece of writing under their surveillance. In this way, we should still be able to use ordinary grids or rubrics.

The use of this powerful tool in student assessment, AI, imposes on us, teachers, the task of carefully thinking about the activities we ask students to carry out and whether the AI outputs do not bring with them prejudiced opinions or actions or that interfere with the well being of students.

On the other hand, as educators, it is also our responsibility to enlight students about the issue of prejudice, eventually, present in AI.

In this particular situation, the algorithm (AI) can be compared with real human assessments, which make it possible to see the student's level of development without additional questions other than evaluating knowledge or evaluating it in a specific area of knowledge.

In other words, we can always take advantage of proposals for activities and tasks, but always in interaction with the teacher's perspective as mediator. Ai is an educational challenge.

In addition to these approaches, integrating human feedback into the assessment process is crucial. While AI is a valuable tool, the nuanced understanding and subjective insights that humans bring are essential. Peer reviews, teacher assessments, and constructive feedback can add a human touch to evaluations. Teaching digital literacy and ethics is another key aspect. Equipping students with the skills to critically evaluate AI-generated content and understanding the potential biases in algorithms fosters a more informed use of technology for academic purposes.

Diversifying assessment methods is also recommended. Instead of solely relying on written assignments, incorporating oral presentations, group projects, and other forms of assessment that may be less susceptible to AI manipulation contributes to a comprehensive evaluation.

Regarding formative assessment strategies, regular check-ins with students provide insights into their understanding, progress, and challenges. This ongoing dialogue enables educators to identify issues early on and offer timely support. Peer collaboration is another effective strategy, fostering a collaborative environment where students can work together on assignments, promoting a sense of community and peer learning.

Exploring adaptive learning platforms that provide personalized feedback based on individual student performance is also beneficial. These platforms can adapt to each student's learning pace and style, enhancing the effectiveness of formative assessment. Additionally, incorporating reflective practices where students assess their own work, set goals for improvement, and track their progress over time encourages metacognition and a deeper understanding of their learning journey. By implementing these strategies, educators can navigate the challenges posed by AI-generated content and algorithmic bias, promoting fair and effective assessment practices in the evolving landscape of education.

As for adapting our assesement practices to students' use of generative AI, some strategies are: asking students to explain some ideas in their text, or assigning references that we are familiar with and students have to use as a base for their text.

I am not afraid, I am pleasently surprised with unexpected possibilities of AI as colaborator, help, learning platform... Use of AI will force education system, and its carriers, to move emphasies from learning facts towards critical thinking, problem solving, social skills... everything that human mind is, still :), are better than AI.

I am math teacher and it is obvious that aproach towards teaching math should move from procedures to understanding, from calculus itself towards problem solving, learning based on researches, critical thinking and so on. The first consequence that I see, for some time, is useles of homework as individual student work in shape that we use in presence. Teacher must change their outcomes and goals in teaching process.

As far as assessment in classroom, techers regular monitoring and data that he collect can improve student learning, and AI can be welcomed just as help but not substitute.

I couldn't agree more with Ivona. I firmly believe that students need to work on essays during school hours, otherwise when at home, the temptation to use AI will be irresistible, as it is for them now to copy from school curriculum-based solution books they can easily find, or making mom and dad to practically write for them. Otherwise, if they work at home , the presentation should be oral, so that even if they have "cheated", oral presentation means incorporating the benefits from their work. (It is like when we miniature- wrote on a tiny cheating paper "skonaki" to hide in our sleeve during the test, but then we realized that writing it down made us learn the topic by heart)

In this perspective, educators need to adapt and flexibilize assessments to accommodate different learning styles, offering a greater variety of assessment formats to allow students to choose methods that best suit their abilities and preferences. Assessments must reflect a variety of perspectives and experiences to avoid cultural or social bias. It is advisable to incorporate diverse materials and examples to enhance fairness, regardless of students' backgrounds. The importance of competency-based assessment over memorization-based assessment should be emphasized. It is also crucial to involve students in the assessment process by participating in the definition of criteria and goals, as well as incorporating self-assessment procedures and self-regulation of learning, encouraging self-reflection and the development of metacognition.

Regarding concerns about algorithmic bias, we need to ensure algorithmic transparency in the formulation and application of assessments, clearly explaining how results are obtained. Human review is essential to mitigate potential algorithmic biases. It is also crucial for students to have the opportunity to contest and question assessment results.

We need to demand that companies providing AI certify their training with diverse and bias-free datasets, as well as ensuring constant monitoring of their algorithms and making continuous adjustments and improvements. Consideration of cultural and social implications in assessment metrics is also necessary to ensure they do not favor a specific group. Educational institutions need to establish ethical policies and practices for the use of algorithms in assessment. Educators, in turn, must create more inclusive, fair, and effective learning environments, adapting to changes in students' learning approaches.

1. Embrace Diversity in Assessment:

Recognize and celebrate the diverse ways students learn by incorporating various assessment methods. This might include traditional exams alongside interactive projects, presentations, or online discussions.

2. Foster Continuous Feedback:

Cultivate a culture of ongoing learning through formative assessments. Regular feedback, whether from peers or self-assessment, becomes a valuable tool for student growth. Leveraging technology for efficient feedback delivery is key.

3. Flexibility is Key:

Acknowledge the different paces at which students learn by offering flexible deadlines and multiple attempts on assessments. This not only accommodates diverse learning schedules but also promotes resilience and improvement.

4. Real-World Problem Solving:

Shift the focus from memorization to real-world problem-solving skills. Encourage open-book or open-resource assessments to mirror the environments students will encounter beyond the classroom.

5. Inclusive Design Matters:

Design assessments with inclusivity in mind. Clear instructions and adherence to universal design principles ensure assessments are accessible to all students, regardless of their background or abilities.

6. Tackling Algorithmic Bias:

Stay vigilant about potential biases in automated grading tools. Regularly review and update algorithms to minimize bias and consider alternative assessment methods when necessary.

7. Transparency and Communication:

Foster an open line of communication with students. Clearly communicate assessment criteria, explain the purpose of assessments, and demystify any algorithmic tools used, ensuring students understand the evaluation process.

8. Continuous Educator Growth:

Invest in professional development opportunities for educators. Training on inclusive and bias-free assessment practices keeps us informed and ensures we're equipped to navigate the evolving educational landscape.

9. Regular Evaluation and Adaptation:

Continuously assess the effectiveness of our assessment methods. Student feedback and performance data will guide adjustments, and staying informed about emerging educational technologies is key.

It is always a challenge to minimize bias in assessment, whether it is performed by a person or by AI. A middle way is probably the best solution, particularly in this early phase of AI development.

The concept of the flipped classroom is becoming more important - teachers should use the valuable time they have with their students in the classrooms (or other learning sites) to encourage face to face communication between students, concentrate on problem - solving tasks, provide valuable formative assessment face to face and present their students various learning methods, including AI assisted learning.

Instead, Art assignments show merciless any mental of physical health issue to the eye of any experienced teacher or psychologist, why AI would not detect that, too. I would not use it for Art assessment. But it's so useful to plan lessons and especially rubrics.

A useful exercise that I think made all the difference in my class was to use it for the discussion of lab reports. A discussion was done in class on the possilbe outcome using ChatGPT for an exemplar experiment that was then regarded. Students had the time to argue the pros and cons and were shown how they could be inspired from what is generated without the need to copy and paste throughout so that they make it their own.

I. The student should argue or present his work orally.

II. In the STEM field, applying the steps of the scientific method reduces the algorithmic bias in education.

Personalized and Diverse Assessments:

Move away from traditional, one-size-fits-all assessments. Tailor assessments to accommodate diverse learning styles and preferences.

Utilize a mix of assessment types, such as project-based assessments, collaborative assignments, and real-world problem-solving tasks.

Continuous Formative Feedback:

Implement continuous formative assessment strategies to provide ongoing feedback, allowing students to understand their progress and make improvements.

Leverage technology for instant feedback through AI-powered tools, promoting timely and targeted interventions.

Transparency and Explainability:

Ensure transparency in the assessment process, providing clear expectations and criteria for evaluation.

Address algorithmic bias concerns by adopting assessment tools with transparent algorithms. Educate students on how AI is used in their assessment and provide explanations for outcomes.

Inclusive Design:

Design assessments that are inclusive and considerate of students with diverse backgrounds, learning needs, and abilities.

Regularly review assessment tools and practices for potential biases, adjusting them to promote fairness and equity.

Training and Awareness:

Provide training for educators on the use of AI in assessment and its potential impact on students.

Foster awareness among students about the presence of AI in assessments, promoting a better understanding of the technology and its role in their learning journey.

Ethical Considerations:

Establish clear ethical guidelines for the use of AI in assessments, ensuring the responsible and unbiased deployment of technology.

Regularly assess the ethical implications of AI-powered assessment tools and make adjustments as needed.

Flexibility and Accessibility:

Allow for flexibility in assessment formats to accommodate various learning environments, including in-person and remote settings.

Ensure that assessments are accessible to all students, considering factors such as language, disabilities, and technological access.

Continuous Evaluation and Improvement:

Continuously evaluate the effectiveness of AI-driven assessments and be open to modifying practices based on feedback and emerging best practices.

Encourage a culture of innovation and experimentation, fostering an environment where educators can adapt their approaches to meet evolving needs.

By incorporating these strategies, educators can harness the benefits of AI in assessments while addressing concerns related to fairness, bias, and effectiveness in student evaluation.

As a teacher, we cannot ignore the potential challenges and opportunities we face with the rise of AI and its impact on the development and delivery of assessments.

By incorporating these strategies into formative assessment practices, educators can create a more balanced and equitable learning environment, empowering students to critically engage with technology and navigate potential biases within AI-driven assessment tools.

Diversify Assessment Methods: Relying on a variety of assessment methods, such as oral presentations, interviews, and practical demonstrations, can help mitigate the impact of AI-generated work and reduce the potential for algorithmic bias to influence outcomes

1

Transparent Assessment Criteria: Clearly communicating assessment criteria to students can help them understand how their work will be evaluated, regardless of whether it is AI-generated or not. This transparency can also help mitigate the potential influence of algorithmic bias on students' work and performance

3

Formative Assessment: Implementing formative assessment strategies, such as providing ongoing feedback and opportunities for revision, can help address concerns related to algorithmic bias. By engaging in a continuous feedback loop, educators can identify and address any potential bias in students' work before it significantly impacts their performance

3

Algorithmic Bias Assessment: In addition to traditional assessment practices, educators can also consider conducting algorithmic bias assessments to evaluate the fairness of AI-generated work. This involves examining the underlying algorithms and data sources to identify and mitigate any potential biases that may impact the assessment process

4

By combining these strategies, educators can adapt their assessment practices to accommodate the evolving landscape of student learning while ensuring fairness and effectiveness.

I'm more on embracing than fight the new tech, so I think we need to find a way to let students openly use AI gererative tools, discussing with them the limits we need to put in their use at school.

The rise of AI tools, such as ChatGPT and other text generators, has posed significant challenges for traditional assessment methods. As students increasingly rely on AI-generated content for essays, projects, and assignments, educators need to rethink how they design assessments to ensure both integrity and effectiveness. Here are some strategies to adapt assessment practices in light of this shift:

1. Incorporating Authentic Assessments

One way to adapt is by designing authentic assessments that require students to demonstrate their knowledge and skills in real-world contexts. Instead of relying solely on traditional essays or tests, educators can ask students to apply concepts in project-based learning, case studies, and problem-solving activities. These types of assessments are harder to outsource to AI tools because they require critical thinking, creativity, and personal input that goes beyond mere content generation​(

CIRCLS

).

2. Utilizing Formative Assessment Strategies

Formative assessment strategies, such as learning journals, classroom discussions, and ongoing reflections, can help address the challenge of AI-generated work. These methods involve continuous monitoring of student learning, allowing teachers to track the development of ideas and detect sudden shifts in writing style that may indicate AI usage. By emphasizing the learning process over the final product, formative assessments can mitigate the impact of algorithmic bias and AI-generated content​(

TechNotes Blog

).

3. Implementing Oral Assessments and Presentations

Oral assessments and live presentations can provide an additional layer of authenticity. Asking students to verbally explain their thought process or present their findings can reveal their depth of understanding, making it difficult to rely on AI-generated content. This approach can be particularly effective in settings where verifying the originality of written assignments is challenging.

4. Exploring AI-Driven Assessment Tools for Detection and Insight

Paradoxically, AI can also be part of the solution. Tools like Turnitin’s AI writing detection or Grammarly’s similarity check can identify AI-generated content by analyzing patterns and discrepancies in writing style. While these tools are not foolproof, they offer a first line of defense against misuse of AI. Additionally, AI can provide insights into student performance patterns, helping to identify cases where algorithmic bias might be influencing student outcomes​(

European School Education Platform

).

5. Adapting Rubrics to Include Process-Based Criteria

Instead of focusing solely on the final product, rubrics should be adapted to assess process-based criteria such as research notes, drafts, and iterative improvements. This ensures that students are evaluated on the basis of their engagement and progression rather than just the end result, making it harder to submit AI-generated work without contributing original thought.

Addressing Algorithmic Bias in Student Performance

Algorithmic bias in AI tools can skew results and misrepresent student abilities, especially for underrepresented groups. For example, language models may produce biased or inaccurate outputs based on incomplete or biased training data. To address this, educators should:

1. Educate Students on Algorithmic Bias: Teach students to recognize and critique bias in AI outputs. This includes understanding how biases in language models can influence their work and learning to cross-reference AI-generated content with reliable sources.

2. Integrate Bias-Detection Tools: AI-powered bias-detection tools can be integrated into the assessment process to flag potentially biased content before it is used in student work. These tools can highlight problematic language or perspectives, making students more aware of the ethical considerations of using AI-generated content.

3. Foster Critical Thinking and AI Literacy: Emphasizing critical thinking and AI literacy is crucial in helping students navigate the use of AI tools. By fostering discussions around the limitations and ethical implications of AI, educators can empower students to use these technologies responsibly.

Conclusion

As AI continues to shape the educational landscape, adapting assessment practices to accommodate AI-generated work and mitigate algorithmic bias is essential. By emphasizing authentic, process-based, and oral assessments, utilizing formative strategies, and integrating AI literacy into the curriculum, educators can create a fair and effective assessment environment that maintains academic integrity and supports student growth.