AI, Cheating and the Crisis in Assessment
Students calls for clearer guidance and fairer access to AI.
Good morning, AI educators.
UK students are calling for stronger guidance and fairer access to AI, while universities face mounting pressure to rethink how they assess learning in the age of ChatGPT.
Together, these stories suggest education is at a crossroads - caught between accelerating adoption and institutional uncertainty.
In this week's news
Jisc’s student AI report calls for curriculum reform and consistent policy across UK education
Universities struggle with AI cheating, unclear policies, and over-reliance on detection tools
Demis Hassabis says teens must become “AI ninjas” to stay competitive
Bridget Phillipson urges educators to embrace AI to boost equity and reduce workload
Wired explores the ethical blur of AI-assisted student work
The Student Perceptions of Artificial Intelligence (AI) Report 2025
UK students are embracing AI tools but are increasingly calling for clearer policies, equitable access, and involvement in shaping how AI is used in education.
Details:
Students use AI daily not just for studying, but also for job prep, mental health, and managing daily life.
Top concerns include employability, unequal access to AI tools, data privacy, and skills loss from over-reliance.
Many want AI literacy embedded into the curriculum and support to prepare for AI-driven job markets.
Jisc recommends consistent institutional policies, support for equitable access, and student involvement in AI strategy.
A balanced approach is urged, preserving human interaction and essential skills like creativity and critical thinking.
Why it matters:
This report offers a crucial insight into how students are adapting to AI and what they expect from their institutions in return. Educators and leaders must respond with clear, inclusive policies that protect equity and prepare learners for a changing workforce. Ignoring these calls could widen digital divides and leave students underprepared in both academic and professional spheres.
Infodocket, May 24, 2025.
‘An existential crisis’: can universities survive ChatGPT?
UK universities are struggling to adapt to the widespread use of AI tools like ChatGPT, leading to concerns over academic integrity, inconsistent policies, and the fundamental purpose of higher education.
Details:
Solicitor Rhys Palmer reports a surge in students seeking legal help over AI-related misconduct accusations, often due to unclear university policies.
Surveys indicate that while most students use AI for non-cheating purposes like proofreading, universities report increasing misconduct cases, particularly at leading institutions like the University of Sheffield.
Universities are belatedly updating AI policies, with approaches varying from allowing AI for research to outright bans, and penalties ranging from warnings to expulsion.
Critics argue that over-reliance on detection software like Turnitin can lead to false accusations, as these tools may flag students unfairly.
Educators express concern over AI's long-term implications for learning and assessment but also acknowledge its potential to enhance education if used responsibly.
Why it matters:
The rapid integration of AI tools in academia challenges traditional teaching and assessment methods, necessitating a re-evaluation of educational practices. Without clear and consistent policies, students risk facing unfair penalties, and the value of a degree may be undermined. Universities must balance technological advancements with keeping academic integrity and equitable access to education.
The Times, May 24, 2025.
AI education news in brief
Teens Should Be Training to Become AI 'Ninjas,' Says Google DeepMind CEO
Demis Hassabis urges teenagers to embrace AI tools to stay competitive, likening AI's impact to that of the internet and smartphones.
May 24, 2025. Business InsiderLet's Talk About ChatGPT and Cheating in the Classroom
A WIRED podcast discusses the ethical dilemmas of AI use in education, highlighting the blurred lines between efficiency and academic dishonesty.
May 24, 2025. WIREDEmbracing AI in Education Will Break Down Barriers to Opportunity
Bridget Phillipson advocates for integrating AI in education to enhance accessibility and provide equal opportunities for all students.
May 22, 2025. LBCAI in Education: Let's Not Miss the Forest for the Trees
Jonathan Becker cautions against overemphasizing AI in education, suggesting a balanced approach that prepares students for broader societal challenges.
May 21, 2025. Virginia MercuryWhat AI Can’t Read: Ambiguities and Silences
Jan Burzlaff discusses the limitations of AI in interpreting complex human experiences, emphasizing the importance of human insight in education.
May 23, 2025. Inside Higher EdShould I Teach My Kids to Use AI?
Anna North and Adam Clark Estes explore the pros and cons of introducing AI tools to children, highlighting the need for parental guidance.
May 22, 2025. VoxAI Takes Over Graduation Name Calls to Fix Mispronunciations, Students React: 'Getting the Ick at My Own Graduation'
Pace University's use of AI to announce graduates' names sparks mixed reactions, raising questions about technology's role in personal milestones.
May 24, 2025. edtechmagazine.com
In other AI news
OpenAI partnered with Jony Ive (designer of the iPhone) to develop AI "companions" and announced plans for a massive data centre in Abu Dhabi, signalling a significant expansion in AI hardware and infrastructure. Axios
Anthropic launched its Claude 4 series, including the advanced Claude Opus 4 model, which can perform complex tasks over extended periods but has raised safety concerns due to its potential for autonomous behaviour. Axios
Aurora, an AI model developed by Professor Paris Perdikaris and Microsoft, demonstrated superior accuracy in weather forecasting compared to traditional methods, particularly for extreme events like Storm Ciarán. The Times
Major newspapers, including the Chicago Sun-Times and the Philadelphia Inquirer, faced criticism for publishing AI-generated summer reading lists featuring fictitious book titles and fabricated quotes, highlighting concerns over AI's role in journalism. The Washington Post
ScienceDaily reported on a study where AI models were tested on emotional intelligence assessments, exploring AI's capability to understand and suggest appropriate behaviour in emotionally charged situations. ScienceDaily
Hi David,
Could you recommend any AI apps that allow us to create conversations between two people speaking as audio files? I'd love to create some examples for listening comprehension practice.