{"id":589063,"date":"2026-08-19T03:08:16","date_gmt":"2026-08-19T03:08:16","guid":{"rendered":"https:\/\/www.newsbeep.com\/il\/589063\/"},"modified":"2026-08-19T03:08:16","modified_gmt":"2026-08-19T03:08:16","slug":"does-ai-watermarking-mean-the-party-is-over-for-cheating-students","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/il\/589063\/","title":{"rendered":"Does \u2018AI-watermarking\u2019 mean the party is over for cheating students?"},"content":{"rendered":"

Concerns about students cheating with AI are rife<\/a> within schools and universities. <\/p>\n

Some lecturers have even expressed fears<\/a> students have been \u201clobotomised by AI\u201d amid reports AI-detection tools<\/a> are patchy and far from fool-proof. Last week, New South Wales moved to ban take-home<\/a> assignments for Year 12 to try to stop AI use in assessments. <\/p>\n

At the same time, a major generative AI system has developed a way to \u201cwatermark\u201d whether or not a text is AI. Developer Anthropic, has announced<\/a> the text output of Claude (one of the five big AI tools) will include a hidden label or watermark indicating the text is AI-generated. <\/p>\n

Files generated by Claude, including Word and PowerPoint, will also contain an additional watermark hidden in the file itself. <\/p>\n

What does this mean for efforts to stamp out cheating? <\/p>\n

Read more:
\n Oral exams are making a comeback to stop AI cheating. But they have their own problems<\/a><\/p>\n

Why a watermark?<\/p>\n

Anthropic took this step after signing up to the European Union\u2019s new AI act<\/a>. The act requires providers of AI systems that create synthetic audio, video, images or text to identify the outputs as AI-generated. <\/p>\n

Other tech companies, such as OpenAI<\/a> (which runs ChatGPT), have developed a text watermarking tool, but not rolled it out publicly. <\/p>\n

The watermarking is not an image superimposed over the text. It is invisibly embedded through the text<\/a>. Readers will need a \u201cdetection mechanism\u201d (yet to be released by Anthropic) to pick up the watermark.<\/p>\n

If this sounds too good to be true, that\u2019s because it is.<\/p>\n

There are still loopholes<\/p>\n

As with AI detection, watermarking can be evaded<\/a> using other AI tools. The absence of a marker also doesn\u2019t guarantee that AI wasn\u2019t used.<\/p>\n

Even Anthropic warns<\/a> their watermarking cannot tell \u201cwhether the text was human-written\u201d. It doesn\u2019t work on small samples of writing. It cannot tell whether AI was used to proofread the work, suggest edits or make improvements. It cannot tell whether a different AI system, such as DeepSeek, wrote the text. <\/p>\n

OpenAI has reportedly<\/a> withheld its text watermarking tool, in part, due to concerns it could negatively impact certain groups, including non-native English speakers.<\/p>\n

AI use has been normalised<\/p>\n

We know AI use is now extremely common among students. The tools are readily available and free to use.<\/p>\n

A 2026 UK survey<\/a> found 95% of undergraduate students use AI, with 94% using it in their assessments.<\/p>\n

Unauthorised AI use is most common in written assessments completed outside of class. Educators have been using AI detection software to try and screen work for cheating. <\/p>\n

A July report<\/a> from software company Turnitin revealed more than 53% of submissions from Australian uni students run through its system used some form of AI.<\/p>\n

However, an increasing number of institutions have abandoned this technology amid concerns about accuracy, bias, procedural fairness<\/a>, and successful legal challenges<\/a> from students overseas.<\/p>\n

What do we need to do instead?<\/p>\n

So it would be unwise for teachers or institutions to rely on watermarking or detection tools to guard against cheating. These technologies can provide a sense of false security. But where does that leave us? <\/p>\n

Some Australian educators are warning<\/a> \u201cdetection-based approaches to integrity are not the way forward\u201d. This is because the focus is on policing whether AI was used, rather than on what the student has learned. <\/p>\n

Or as University of Queensland educational psychology expert Jason Lodge<\/a> has also argued<\/a>, we need to:<\/p>\n

stop looking for evidence that students are using these tools to cheat and shift our emphasis to looking for evidence that learning has occurred.<\/p>\n

What happens now?<\/p>\n

Advice on alternatives to the \u201cdetect and punish\u201d approach has been steadily growing over the past three years. Some of this is via the higher education regulator<\/a>. It notes:<\/p>\n

we need to prepare students to engage actively, responsibly and ethically in a society where AI is everywhere<\/p>\n

understanding what students have learned requires different types of assessment that are fair, inclusive and suited to the situation.<\/p>\n

Concerns about cheating and AI-use have also resulted in calls to return to in-person assessments. But these aren\u2019t workable<\/a> for those who study fully online. They also don\u2019t prepare students to engage with AI. <\/p>\n

The higher education regulator has also outlined<\/a> how assessments could change in practice. This includes assessments that are connected and build upon each other throughout a whole degree, increasing in complexity. It also includes at least one \u201csecure\u201d assessment task in every unit or subject across a degree program. <\/p>\n

Secure tasks can include oral presentations, in-class tasks and supervised practical demonstrations. They must also include mechanisms that verify the person completing the assessment is the enrolled student.<\/p>\n","protected":false},"excerpt":{"rendered":"Concerns about students cheating with AI are rife within schools and universities. Some lecturers have even expressed fears…\n","protected":false},"author":2,"featured_media":589064,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[345,343,344,85,46,125],"class_list":["post-589063","post","type-post","status-publish","format-standard","has-post-thumbnail","category-artificial-intelligence","tag-ai","tag-artificial-intelligence","tag-artificialintelligence","tag-il","tag-israel","tag-technology"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/il\/wp-json\/wp\/v2\/posts\/589063","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/il\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/il\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/il\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/il\/wp-json\/wp\/v2\/comments?post=589063"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/il\/wp-json\/wp\/v2\/posts\/589063\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/il\/wp-json\/wp\/v2\/media\/589064"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/il\/wp-json\/wp\/v2\/media?parent=589063"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/il\/wp-json\/wp\/v2\/categories?post=589063"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/il\/wp-json\/wp\/v2\/tags?post=589063"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}