This keynote gives an overview of why and how AI tools for assessment purposes can be used. One part of the presentation covers AI-based Proctoring Systems, another part moves closer into AI tools for assessments, and a last part looks at university guidelines, ethical considerations, some pedagogical options to embed AI tools for students while they work on projects, and some AI tool resources.
Keynote AI assessment tools: online exams and tools.pptx
1. Educational assessment in an AI world
Inge de Waard. 16 November 2023. Virtually between Graz (Austria) and Maribor (Slovenia).
All pictures created with Bing Enterprise (= Dall-E 3)
2. How do we Teachers
cope in times of AI?
• Teacher tasks in times of AI: learning,
teaching, research, university admin…
• Using AI in combination with human
analysis (critical thinking, peer
knowledge building…)
3. What is AI?
Everyone and no
one knows
AI tools ? Software applications
that use artificial intelligence (AI)
techniques:
o natural language processing
o machine learning
o deep learning
But at the end … us teachers need
to teach, and learn with it.
4. Is AI in Education
here to stay?
Do you know of any type of technology
that was invented, but then stopped due
to reasons of having a negative influence
on our planet/people?
If so, please put it as a short answer in
the chat
5. The European University Association
promotes responsible use
While GenAI is increasingly implemented, the EUA
made a statement (with documentation) to move
towards a consistent, transparent responsible use of
AI (the EUA took this position on 14 Feb 2023)
6. A good start …
sets the stage
Are you using AI to create or
support assignments?
• Yes = thumbs up
7. Using AI tools
within University
education
AI tools can be used for various
purposes in university education:
• enhancing learning (EU: TutorAI,
MathsGee),
• supporting research: Elicit,
SemanticScholar = free
• facilitating communication: EU
DeepL (now integration for
Windows),
• And: enhancing assessment
options.
8. Impact of AI tools
on University
Education
AI tools pose challenges and risks:
o academic integrity
o quality assurance
o ethical standards in university
education.
10. 10
Topics
Using and understanding current
AI assessment options
Current trends in AI tools
Online exams Proctoring tools
AI assessment tools scope
Benefits and drawbacks
Guidelines and best practices
11. Current
trends:
AI tools in
education?
Scaling: university needs to expand, as
well as attract more students or
learners / alumni / work closer with
business
Student support: personalized
learning, tutoring support…
Expand online learner reach
16. Some AIPS
examples
Constructor Proctor. An online proctoring tool
to detect mobile phone, book, laptop, multiple
people, key press, and tab switch during the
exam (EU).
Proview: A proctoring SDK for Android that
integrates with various assessment platforms
and uses AI to detect cheating behavior.
Proctoring-AI: A GIThub software for automatic
monitoring in online proctoring that uses
computer vision and natural language
processing to analyze the video and audio of
the test-taker.
18. Considerations using a proctoring tool
Compatibility with the LMS / platform and student devices.
Inform the students about the proctoring tool and its features.
Provide clear instructions and expectations for the online exam,
such as the testing rules, the duration, the allowed material.
Have the students sign an honesty statement prior to being
allowed to see the test. This can help to promote academic
integrity and remind students of the consequences of cheating
behavior.
Use a lockdown browser to minimize external help. A lockdown
browser is a software that prevents the students from accessing
other applications, websites, or files during the exam (no copy-
paste, print …).
Use question pools so no two exams are alike. Question pools are
sets of questions that are randomly selected and presented to the
students. It reduces chances of copying or sharing answers among
students.
21. What makes a good assessment?
• Covering knowledge or experiences provided
• Including multiple levels of understanding/implementation
• Ranging in difficulty
• Providing a good granularity for grading
A nice tight combination with the content provided.
22. Assessment combined
with course creation
• Coursefactory AI co-pilot: build courses, while integrating
assessments/quizzes in the same effort (e.g. multiple
choice, fill in the blancs, open ended questions and
evaluation)
• Nolej.ai to create interactive material, from your own
databanks, while optimizing the alignment of content
and assessments.
23. AI assessment tools for
written assignments
• EssayGrader - Grade essays with the power of AI
(including a summarizer to get the overall gist of an
essay.)
• QuestionPaper AI to create a set of questions for
each student’s story, based on the content, style,
and theme of the story. The questions were
designed to test the students’ comprehension,
analysis, and evaluation of their own work, as well
as their creativity and originality.
24. Questions on AI tools?
And a quick stretch for those of us who have been
seated the whole time…
26. Students can’t use AI tools during tests
These are assessments where the use of AI tools
would compromise the learning outcomes, the validity,
or the fairness of the assessment.
For example: multiple choice questions, short
answer questions, or essays that require
students to demonstrate their own knowledge,
understanding, or critical thinking skills.
27. Assistive AI assessment use
AI tools can be used in an assistive role. These are assessments where
the use of AI tools can support or enhance the students’ learning
process, but not replace their own work or judgment.
For example, AI tools that can help with idea generation (e.g. Bing
Chat Enterprise), planning, data mapping (e.g. ChartGPT), referencing,
or feedback. Or chatbots that assist the students learning progress.
28. Integral use for assessments
For example, AI tools that can help with data analysis, problem solving,
creativity, or innovation.
AI has an integral role. These are assessments where the use of AI tools is
expected or required, and where students need to demonstrate their ability
to use AI tools effectively and critically.
30. Benefits of
using AI
assessment
tools
Improving the efficiency and scalability of assessment
(reducing the workload and time required for
marking, grading, or providing feedback).
Enhancing the diversity and flexibility of assessment
(more personalised, adaptive, and formative
assessment methods and formats).
31. Drawbacks
of using AI
assessments
tools
Threatening the integrity and quality of assessment
Raising ethical and legal issues, by involving the use of
personal data, intellectual property, or sensitive
information, and by requiring the compliance with data
protection, privacy, or consent regulations and policies.
Creating digital divides and inequalities, by requiring
access to technology, internet, or digital literacy, and by
creating gaps or biases between different groups of
students, staff, or institutions.
33. Integrating AI in student activities
Creative writing task. The teacher assigns a creative writing task to the students, asking them to write
a short story based on a prompt. The teacher also asks the students to share their writing process:
how they came up with the idea, what challenges they faced, and what feedback they received from
peers (in case of peer evaluation as part of the writing process).
Activity plus side notes. The teacher asks students to answer a set of questions regarding the process
in a separate document from a writing/production activity, and submit both the story/product and
the answers regarding the process to the assessment. The questions include citing any sources or
tools the students used for their writing, product realization, … including AI tools.
Plagiarism checkers. The teacher can use plagiarism checkers and readability analyzers, to verify the
authenticity and quality of the stories. In this case it is crucial to share that fact in advance with the
students to create an environment of transparency. (Be warned, students are smart)
Ethics discussion. Any learning/teaching activity including an AI tool, provides a good opportunity to
discuss ethical and academic implications of presenting AI-generated content as one’s own.
34. AI tools as learning assistants for students
POE chatbot for many
platforms and purposes.
AI-mentor: mostly aimed
at skills – personal and
professional.
Quillbot (re)writing in
style and timbre
AI-tutor: UK-based math
AI tutor.
38. Educate students on ethical AI use
Ethical guidelines on the use of artificial intelligence and data in teaching
and learning for educators (EU)
The guidelines provide practical advice and examples on how to
integrate the effective and ethical use of AI and data into school
education
How to develop the relevant competences among teachers/ students.
Teachers can use AI tools as a catalyst to foster dialogue and debate
among students, as well as with other stakeholders, such as parents,
experts, or policymakers.
Teachers can use AI tools as a model to demonstrate and exemplify the
ethical principles and values that should guide the use of AI in
education and beyond
Have you heard or tested https://roft.io/ (identify fake or real text, it is a
game)
39. Guidelines
Students and staff need to follow and adhere to the policies and guidelines of
academic integrity, quality assurance, and ethical standards, and to report and address
any cases of misuse or misconduct.
Example guidelines
• University of Greenwich (UK) https://www.gre.ac.uk/articles/public-
relations/guidance-on-the-use-of-artificial-intelligence-ai
• KULeuven (Belgium)
https://www.kuleuven.be/english/education/student/educational-tools/generative-
artificial-intelligence
41. Conclusion
•
AI tools can help
But also pose some risks
Use them with caution
Source: Bing Chat enterprise haiku
o AI tools have the potential to transform the way
assessment is designed, delivered, and evaluated.
o The implementation comes with challenges and
risks for academic integrity, quality assurance, and
ethical standards
o A clear university wide regulation for all
stakeholders is needed to ensure their appropriate
and effective use.
o Students and staff need to embrace and engage
with AI tools in assessment, as an opportunity to
enhance their learning, research, and innovation.
42. Useful resources for AI use
• How to prevent hallucinations. Zapier article (useful when
using AI tools for content creation, Sep 2023)
• OpenRefine. Free open-source tool for data cleaning
(useful for students)
• Timeline of AI up until now. A nice interactive chart and
descriptions.
• Learning prompts (or buying them):
https://promptbase.com/
Two AI repositories (nice to find latest and highest evaluated
tools:
• Futurepedia, an AI tool repository
• There’s an AI for That, another AI tool repository
43. www.innoenergy.com
EIT InnoEnergy
Kennispoort 6th floor
John F. Kennedylaan 2
5612 AB Eindhoven
The Netherlands
Info@innoenergy.com
Inge de Waard – Questions?
• LinkedIn: https://www.linkedin.com/in/ingedewaard/
• Inge.dewaard@innoenergy.com
Editor's Notes
, to generate or manipulate text, images, audio, or other types of data.
Natural language processing (NLP) is a branch of artificial intelligence (AI) that deals with the interaction between computers and human languages, such as English, Chinese, or Arabic. NLP enables computers to understand, process, and generate natural language texts, such as emails, tweets, articles, or books. NLP also enables computers to perform tasks that involve natural language, such as language translation, sentiment analysis, text summarization, question answering, or chatbot development. NLP uses various techniques and methods, such as rule-based systems, statistical methods, or machine learning algorithms, to analyze and manipulate natural language data12.
Machine learning (ML) is a subfield of AI that focuses on creating systems that can learn from data and improve their performance without being explicitly programmed. ML enables computers to discover patterns, make predictions, and perform tasks that are too complex or tedious for humans to do manually. ML uses various algorithms and models, such as decision trees, support vector machines, or neural networks, to learn from data and make decisions or recommendations13.
Deep learning (DL) is a subfield of ML that uses artificial neural networks to model complex functions and processes. Neural networks are composed of layers of interconnected nodes that mimic the structure and function of biological neurons in the brain. DL enables computers to perform tasks that require high-level abstraction and representation, such as image recognition, speech synthesis, natural language generation, or face detection. DL uses various architectures and techniques, such as convolutional neural networks, recurrent neural networks, or generative adversarial networks, to learn from large amounts of data and produce realistic and accurate outputs14.
Introduction – AI in general (including GenAI was given in prior presentations, so I will not go into the different aspects of AI in general but focus on AI assessments.
EU based AI tools
TutorAI, a web-based platform that uses AI to provide personalized learning modules on various topics. You can enter a topic and the AI will return several lessons, examples, and quizzes for you to learn from. TutorAI was developed by Easy With AI, a company based in London, UK1.
MathsGee, an online platform that offers interactive math courses and exercises for students of all levels. MathsGee uses AI to adapt the content and difficulty to the student’s needs and progress. MathsGee was created by Edzai Zvobwo, a Zimbabwean entrepreneur who is based in Geneva, Switzerland2.
DeepL is an AI-driven translation tool that was invented by DeepL GmbH, a company based in Cologne, Germany1. DeepL GmbH was founded in 2009 by Gereon Frahling and Leonard Fink, who were former Google employees and researchers2.
Specifically related to the implications and recommendations for students and staff.
The presentation is structured as follows: first, I will review some examples of AI tools that can be used in assessment, and categorise them according to their role and function. Second, I will analyse the benefits and drawbacks of using AI tools in assessment, and highlight some of the key issues and dilemmas that need to be addressed. Third, I will propose some guidelines and best practices for using AI tools in assessment, and suggest some ways to foster academic integrity and quality in the age of AI.
Facial Recognition: AI proctoring software uses facial recognition technology to verify the identity of the test-taker. It captures the test-taker's face during the authentication process and then continuously monitors the face throughout the exam to ensure the same person is taking the test.
- Browser Lockdown: To prevent cheating, the AI proctoring software may lock down the test-taker's browser, limiting their ability to access external websites or applications during the exam.
- Screen Monitoring: AI algorithms monitor the test-taker's screen to detect any suspicious activities, such as switching between windows, opening unauthorized applications, or attempting to copy-paste.
- Keystroke Analysis: Some AI proctoring systems analyze the test-taker's keystrokes to detect unusual typing patterns that may indicate cheating or the presence of a second person.
- Audio Analysis: The software may also monitor audio to detect any unusual sounds or voices in the background that could indicate unauthorized communication.
- Real-time Monitoring: AI proctoring tools provide real-time monitoring, where human proctors or instructors can intervene if they detect any suspicious behavior through the AI system's alerts.
- Recorded Sessions: In addition to real-time monitoring, AI proctoring systems often record the entire exam session. These recordings can be reviewed later for further analysis and evidence if any irregularities are suspected.
.
Guidelines and best practices for using AI tools in assessment: In this slide, I provide some recommendations and suggestions for using AI tools in assessment, and outline some strategies and actions to promote academic integrity and quality in the age of AI.
Students and staff need to be trained and supported on how to use AI tools in assessment, and to develop the skills and competencies to evaluate, critique, and improve their use and outputs.
Conclusion:
AI tools have the potential to transform the way assessment is designed, delivered, and evaluated.
The implementation comes with challenges and risks for academic integrity, quality assurance, and ethical standards
A clear university wide regulation for all stakeholders is needed to ensure their appropriate and effective use.
Students and staff need to embrace and engage with AI tools in assessment, as an opportunity to enhance their learning, research, and innovation.
Rather pragmatic Haiku provided by Bing Enterprise Chat on the subject of AI tools
Can this gif be replaced by a dall-E 3 generated gif?