This post is the third in a series of nine articles updating the “Teaching AI Ethics” resources published in 2023. In the original series, I presented nine areas of ethical concern based on research into artificial intelligence, including the growing technologies of generative artificial intelligence.
In the original post on Truth I split my attention between misinformation and academic integrity as two types of “truth” that we needed to attend to when teaching AI ethics. Since 2023, “cheating” with AI has continued to be an issue, but despite the fact that the media seems to have only just caught onto it, most educators I speak to are already tired of the discussion and ready to move on to actions.
In this updated article, I’m going to dedicate much more time to misinformation, deepfakes, and the inherent issues of using LLMs to generate truth. I will still discuss academic misconduct, but the conversation will shift towards ideas of postplagiarism and how we might move beyond the “cheating” conversation in schools and universities.
For the 2023 articles, which include still-relevant case studies and teaching ideas, check out the original series of posts:
By design, large language models are incapable of truth. We’ve seen this in action since the release of ChatGPT, which initially had no internet connection, no way to upload or attach files, and a tendency to “hallucinate”.
That word, “hallucination”, has become the de facto term for what language models do when they make stuff up. But it is not an error, or a quirk: It is a feature of the way these things are built. Large language models rely on statistical matching and probability. The training data is only used to create the rules governing the languages the model can produce. There is no epistemic grounding, and no “truth” left in the system once trained.
Since 2023, most models have relied on a combination of internet access, file upload, retrieval augmented generation (RAG), and reinforcement learning from human feedback to mitigate rates of hallucinations, but none of these methods have fully guarded against the large language model producing factually incorrect information. Hallucinations are a feature, not a bug. I wrote about the technical composition of chatbots and why they “make sense, not words” in this post:
Even more problematically, although AI can now cite sources from the internet or use processes like RAG to access internal documents, this only increases the perceived credibility, making it harder to spot misinformation. The automation bias, a tendency of humans to believe machine-driven responses, means that people tend to believe what computers tell them.
Unfortunately, discourse on social media often compounds the problem, with AI influencers suggesting that hallucination rates are dropping when, out in the wild, the opposite seems to be true.
Just last week at the time of writing, the US government’s “Make America Healthy Again” in the US clearly contained content generated by OpenAI’s ChatGPT. The original release of the report carried telltale markup of ChatGPT-generated citations and many of the citations were inaccurate, misleading, or pointed to the wrong papers. The issue was reported on in The Washington Post, just prior to the paper being taken down and amended.
Even medicine – a field where truth and accuracy are surely paramount – there have been issues with AI. For example, this highly successful paper detailed ways that language model-based technologies can be used to predict the function of unknown enzymes. However, as Rachel Thomas pointed out in this article, “The Transformer model in the Nature paper made hundreds of “novel” predictions that are almost certainly erroneous.” Thomas goes on to describe that a counter-paper pointing out the inaccuracies, published on the preprint platform bioRxiv, received much less attention than the 22,000 views-and-counting of the Nature article.
Whatever the AI pundits tell you, it’s not about whether more advanced LLMs or reasoning models like ChatGPT o3 Anthropic’s Claude Opus 4 reduce hallucinations, it is the way in which these technologies are actually used by people in high-stakes situations like law, politics, medicine, and education.
Deepfakes and synthetic media
In 2023, I also started writing about what I termed “digital plastic” – the creation of synthetic multimodal texts that I cautioned would swamp the internet over the next few years. At the time, image, audio, and video generation were still fairly nascent and not anywhere near as convincing as text generation. But in the two years since the original Teaching AI Ethics series, that has changed. The truthfulness of audiovisual content online can no longer be taken for granted (not that it ever could, really).
The creation of incredibly convincing AI-generated images is now trivial and possible through ubiquitous open-source technologies with very few restrictions or guardrails. Regulations and laws have been slow to keep up, though in Australia, at least, there have been changes regarding the creation of non-consensual, explicit deepfakes.
Deepfakes are particularly troubling when discussing artificial intelligence and truth, since it is now trivial to create convincing and entirely untrue videos of people saying and doing things they never did or said. 99% of deepfakes circulated online are nonconsensual and explicit, and 98% of those images are of women: this is gender based abuse. These deepfakes are generally images but increasingly include video. While in 2023 this was already becoming problematic for celebrities, advances in these technologies mean that everyone is potentially vulnerable.
Deepfakes can also be used in political misinformation and deliberate reputational damage. We have already seen the deliberate sharing of deepfakes by politicians and world leaders on platforms like X, where moderation has all but disappeared. AI-generated websites, social media accounts, and bots are ubiquitous on most major platforms, generating and spreading misinformation and disinformation. Some of these are naive hallucinations, others deliberately antagonistic and harmful.
Because of the rise of these technologies, major developers have made efforts to secure their AI-generated content with watermarking and provenance standards like C2PA. These approaches identify synthetic media, often applying a content credential or metadata which articulates what platform was used to create the image, video, or audio.
While this legitimises content created on those platforms, allowing users to transparently label their content as AI-generated, it will not stop others from generating harmful deepfakes and synthetic media because they can simply use open-source equivalents which lack those credentials. Authenticity signals are part of the equation, but strong regulation is needed with heavy penalties for people creating and sharing harmful and untruthful content.
In education, one thing we can do right now is to show educators and students just how difficult it already is to spot real from fake. In 2024, I released a little game which has now been played by over 100,000 people, including hundreds of educators. In my face-to-face and online sessions, the average score is 5 or 6 out of 10. It is already harder than you think to spot deep fakes.
As well as creating deepfakes and untruthful content, artificial intelligence contributes to the sharing and distribution of that content. Social media has played a huge role in elections for over a decade now. Many will remember the Cambridge Analytica scandals of Facebook, where collection of user data was implicated in manipulating the voting preferences of voters in the run-up to the 2016 US elections. This kind of disinformation can also be seen in more innocuous places such as the comment threads on YouTube videos, or the insinuation of AI chatbots into Facebook groups.
Artificial intelligence can be deployed to deliberately boost or suppress posts, including those which have been generated by artificial intelligence in the first place, to sway public, political, and social discourse in one way or another, or by advertising companies to sway public opinion on products and services.
Post-plagiarism and moving beyond cheating
All of this contributes to my thoughts on why we need to move discussions in education beyond cheating or catching students using artificial intelligence. Some discussions of academic integrity, for example, exclude artificial intelligence entirely from educational contexts, which makes it very difficult for students and educators to understand the full implications of these technologies in society.
Focusing on whether students using AI technologies are “cheating tools” also obfuscates a more important question of whether we can help students to learn to use artificial intelligence in ways which are truthful, accurate, and reliable.
As I demonstrated above, it is not enough anymore to teach students to try to detect AI-generated fake news: it is already impossible to do so with 100% certainty. Similarly, there is no totally viable way to guarantee that anything produced by a student in the context of digital texts is their own work.
But if detection tools don’t work, and if it’s not possible for an educator to spot the use of generative AI by eye, what does that mean for educating students in a digital context? Students deserve to conduct their work in an environment which reflects the world beyond the classroom. There have to be options for students who require online education. And teachers deserve to be able to teach – including online – without constantly looking over their shoulders or chasing academic integrity concerns.
As I’m writing this article, I’m of the opinion that the most secure form of assessment is what we would call “Level one”, or what others, might call “Lane one” assessments: supervised, in-person, technology-free assessments of student learning, which may or may not be conducted under examination conditions, but which, through both technological and physical means, limit access to artificial intelligence.
In the future, though, even these methods will be increasingly vulnerable. A student who really, really wants to cheat will be able to do so in increasingly sophisticated ways which involve, for example, wearable AI technologies. Of course, the students who have access to those kinds of technologies will tend to be wealthier, and more (digitally) literate.
The point that I am making is this: we cannot solely focus our energy and attention on assuring “truth”. We need philosophies in education which encourage and reward truthful behaviour.
Sarah Elaine Eaton’s post-plagiarism is one such philosophy which I think has important implications for truth in education.
Hybrid writing, co-created by human and artificial intelligence together is becoming prevalent. Soon it will be the norm. Trying to determine where the human ends and where the artificial intelligence begins is pointless and futile.
Human Creativity is Enhanced
Human creativity is enhanced, not threatened by artificial intelligence. Humans can be inspired and inspire others. Humans may even be inspired by artificial intelligence, but our ability to imagine, inspire, and create remains boundless and inexhaustible.
Language Barriers Disappear
One’s first language will begin to matter less and less as tools become available for humans to understand each other in countless languages.
Humans can Relinquish Control, but not Responsibility
Humans can retain control over what they write, but they can also relinquish control to artificial intelligence tools if they choose. Although humans can relinquish control, they do not relinquish responsibility for what is written. Humans can – and must – remain accountable for fact-checking, verification procedures, and truth-telling. Humans are also responsible for how AI-tools are developed.
Attribution Remains Important
It always has been, and always will be, appropriate and desirable to appreciate, admire, and respect our teachers, mentors, and guides. Humans learn in community with one another, even when they are learning alone. Citing, referencing, and attribution remain important skills.
Historical Definitions of Plagiarism No Longer Apply
Historical definitions of plagiarism will not be rewritten because of artificial intelligence; they will be transcended. Policy definitions can – and must – adapt.
From my experience in both K-12 and higher education contexts, there’s a lot to be said in favour of the six tenets of post-plagiarism, particularly its focus on transparency and shared accountability. There will always be students who, for whatever reason, break our trust. But that is by no means all students, and we cannot afford to let this technology, inherently untrustworthy as it is, erode mutual respect between educators and learners.
To get to the truth of whether a student understands what they have learned will be a difficult task, but a worthwhile one. It will include a blend of assessments which permit and exclude artificial intelligence. It cannot do so at the expense of accessibility and inclusion, or these assessments themselves would be invalid. This is the work of education.
Teaching Truth and AI Across the Curriculum
In the original 2023 collection, each article ended with a selection of ideas for teaching the issue in the context of existing curriculum areas. These 2025 updates will similarly align concepts from the articles to standards from typical curricula across the world, and in particular the Australian, UK, US, and IB curricula. For the readers teaching in Higher Education, these examples will also be suitable across a wide range of disciplines.
English
In the English curriculum, students analyse how language constructs meaning, truth, and credibility. With AI tools increasingly producing persuasive texts and opinion pieces, this is an ideal space to explore how generative tools shape narratives. Students might ask “Can AI-generated arguments be trusted?”, “How does AI mimic authorial voice to create a sense of truth?”, or “What rhetorical strategies make false claims produced by AI seem convincing?” These activities connect to media analysis, persuasive writing, and critical discourse study.
Humanities (History / Civics / Geography)
Humanities subjects centre on source reliability, perspective, and representation—areas increasingly complicated by synthetic media. Students might investigate “What happens when AI rewrites history?”, “How can deep fakes influence public trust in political institutions?”, or “How do algorithms impact the information we see—and don’t see—about global events?” These lessons align with source evaluation, historiography, digital civics, and ethical information use.
Science subjects teach students how to evaluate evidence and peer-reviewed claims—skills essential in an era of AI-generated misinformation. Students might ask “How can we tell scientific fact from AI-generated pseudoscience?”, “What happens when AI produces research summaries or citations that aren’t real?”, or “How do hallucinations in large language models affect public understanding of science?” Projects could include analysing AI-generated scientific texts, comparing summaries to source articles, or testing citation credibility.
Digital Technologies / Computer Science
In digital technologies, students can study the architecture of systems that generate, amplify, or suppress information. This aligns with questions like “How do algorithms prioritise certain truths over others?”, “Can we train AI to detect deep fakes—or is it part of the problem?”, or “What ethical responsibilities do developers have when designing tools that could mislead?” Lessons might include building simple classifiers, auditing language model outputs, or exploring watermarking and provenance tools.
Mathematics
Mathematics offers a lens for analysing the claims made by and about AI, especially through data and statistics. Students could ask “What biases in training data affect how AI models ‘learn’ truth?”, or “How can probabilistic outputs from AI be misread as definitive claims?” These tasks align with interpreting data sets, evaluating statistical reasoning, and questioning the perceived objectivity of numbers.
The Arts (Visual Arts / Media / Performing Arts)
In the Arts, students explore representation, symbolism, and the power of visual and performative language. With AI now able to generate hyperrealistic images and synthetic voices, this is a space to ask “How does AI blur the line between real and fake?”, “What responsibilities do artists have when using AI to depict ‘truth’?”, or “How can we spot manipulated or generated media in art and performance?” These questions support critical visual literacy, digital media production, and ethical reflection on creative technologies.
Languages (Other than English)
In language learning, students examine meaning, nuance, and cultural context—areas where AI translations and outputs can distort or misrepresent truth. Students might explore “How trustworthy are AI-generated translations of culturally significant texts?”, “What happens when nuance and idiom are lost in machine translation?”, or “Can we trust AI to represent diverse voices and perspectives accurately?” These tasks encourage cross-cultural understanding, linguistic accuracy, and discussion around the limitations of automated language systems.
Note on the images: In the 2023 version of Teaching AI Ethics I generated images in Midjourney. This time around, I have sourced images from https://betterimagesofai.org/. I still use AI image generators, but due to the environmental concerns and the contentious copyright issues discussed in these articles, I am more conscious of my use. Better Images of AI includes an excellent range of photos, illustrations, and digital artworks which have been generously licensed for commercial and non-commercial use.
All articles in this series are released under a CC BY NC SA 4.0 license.
Subscribe to the mailing list for updates, resources, and offers
As internet search gets consumed by AI, it’s more important than ever for audiences to directly subscribe to authors. Mailing list subscribers get a weekly digest of the articles and resources on this blog, plus early access and discounts to online courses and materials. Unsubscribe any time.
Want to learn more about GenAI professional development and advisory services, or just have questions or comments? Get in touch:
[…] This is an updated post in the Teaching AI Ethics series, originally published in 2023. Given the explosive developments in AI and copyright over the past two years – including major court cases, government decisions, and the first billion-dollar settlements – it felt essential to revisit this intermediate-level ethical concern. For the previous updated post on Truth, click here. […]
Leave a Reply