Academic misconduct is nothing new. It’s just the means of cheating that change. As educators, it’s important to keep up with technological trends that facilitate misconduct, for effective detection and correction.

In this article, we turn our spotlight onto an emerging threat – AI-to-human rewriter services – and explore effective strategies to safeguard assessments in the face of this evolving challenge.

Why do essay mills remain a challenge for academic integrity?

For years, essay mills have met the demand for outsourcing essay writing, offering tailored assignments that match students’ grade aspirations and level of language proficiency.

These outsourced essays make it difficult to detect misconduct, with 67% of instructors saying they may not act on suspicions of contract cheating due to insufficient evidence to support their claims.

Many essay mills are organized criminal enterprises with ethics to match. Contact with these service providers, while superficially innocuous, puts students at risk of extortion and even poses a risk to national security.

To protect vulnerable students from exploitation by essay mills and to safeguard academic integrity, many countries have enacted laws banning the offering and advertising of these services. In the UK, essay mills have been considered illegal entities since 2022 and Internet service providers are prohibited from carrying their adverts. This follows similar action in Australia the year before.

However, it hasn’t solved the problem, simply moved it. In Australia, international students report being bombarded by ads through private messaging services and email (The Guardian, Lures and violent threats: old school cheating still rampant at Australian universities, even as AI rises).

And, now, the rise of generative AI adds a new dimension to academic misconduct, making it free and relatively easy to churn out assignments.

Is AI-facilitated contract cheating ‘better’ than traditional contract cheating?

Some commentators posit that cheating via ChatGPT is preferable to contract cheating as it requires some subject knowledge and involves less contact with malefactors.

Many essay mills are organized criminal enterprises with ethics to match. Contact with these service providers, while superficially innocuous, puts students at risk of extortion and even poses a risk to national security.

They also argue that using generative AI does at least test students’ knowledge of their subject – to create meaningful prompts and assess the accuracy of answers – which is an improvement on simply outsourcing the entire process.

Furthermore, the distinctive patterns of AI-generated content and the sophistication of plagiarism detectors mean that this form of cheating can be more easily identified, allowing for timely corrective interventions.

However, these arguments assume that contract cheating and AI-based misconduct are mutually exclusive. But with essay mills’ business models being disrupted by gen-AI, they’re pivoting to a new approach that supports students as they look to humanize AI.

What are AI-to-human rewriter services?

As the name suggests, AI-to-human-rewriter services receive AI-written content from paying individuals, then edit it to humanize the content and make it less obviously AI-generated.

The purpose of this is to improve its quality, make it more original, or adapt it to a specific style or voice. This, in turn, reduces the risk of plagiarism detection, both from human readers and academic integrity tools.

Essay mills are now targeting students through social media platforms, offering ‘undetectable’, ‘humanized’ assignments that evade plagiarism tools.

AI paraphrasing vs AI-to-human rewriters – what’s the difference

There are important distinctions when it comes to editing AI-generated content. The two most common methods are AI paraphrasing and AI-to-human rewriting. But what’s the difference?

What is the difference between AI paraphrasing and AI-to-human rewriting?

AI paraphrasing tools use algorithms to reword and rephrase existing text while preserving its original meaning. The resulting content can closely mirror the original structure and ideas but with different wording. AI writing detection tools can detect AI-generated content that may have been paraphrased using a word spinner/AI paraphrasing tool.

AI-to-human rewriting is harder to detect as this content does not necessarily match the structure and patterns of the source content. The final text is more likely to reflect a unique style and can be harder to detect as AI-generated.

Why are students turning to AI-to-human rewriter services?

Students are using AI-to-human rewriter services for a number of reasons.

Firstly, there’s the question of why students are driven to contract cheating and misconduct in the first place, regardless of the method. This can be down to:

• Academic pressure and career aspirations
• Fear of failure and familial expectations
• Financial pressure and the time constraints of earning while learning

Next, there's the question of why this particular form of cheating?

Many students are already using AI to support their study, unaware of the negative ethical and learning implications of its use.

Faced with the risk of detection of wholly AI-generated assignments – with many being aggressively targetted by ads for ‘undetectable AI’ essay services – they’re turning to AI-humanizer services to reduce this risk. Like essay mills, AI-to-human rewrite services can be tailored to match individuals’ academic ability, language mastery, and past performance, which makes detection and discovery less likely.

Furthermore, as discussed above, using AI does require a certain degree of subject knowledge and language proficiency. For some students, AI-to-human rewriter services remove even this barrier to academic misconduct.

What are the implications of AI-to-human rewriter services for academic integrity policies?

So-called ‘undetectable AI’ poses a significant challenge for educational institutions. Educators need to root out misconduct in order to:

• Improve individual academic outcomes and support students to success
• Protect their institutional reputation and the value of their qualifications
• Safeguard the reputation of graduates who have earned their degrees legitimately
• Create graduates who are ready to make an economic impact nationally and globally

By humanizing AI, AI-to-human rewriter services make it harder for institutions to identify student misconduct and reduce their ability to offer appropriate interventions to students.

While AI-enhanced essay mills are on the rise, research into contract cheating law shows academic integrity policies are failing to address the problem adequately, putting their reputation in jeopardy.

AI-to-human rewriter services potentially exploit loopholes – in institutional policies and the law – that don’t specifically address humanized AI content. This can make it harder to pursue disciplinary processes and protect institutional reputation.

Action is needed to close the loophole and make institutional policy on this form of contract cheating clear. However, lawmakers and educators need to be careful not to create overly zealous policies that ‘bring bone fide general purpose AI systems into scope unnecessarily’ the same research asserts.

How can institutions respond to AI-to-human rewriter services?

AI-to-human rewrite services continue the debate on the ethics of student use of AI, particularly in a landscape where 60% of future jobs will be impacted by AI. Some may argue that leveraging AI is a critical skill for tomorrow’s workforce.

However, it’s clear that AI-to-human rewriter services – like traditional essay mills – remove any of the critical thinking required to complete assignments. Plus, they are more likely to evade detection than fully AI-generated content.

As a result, they are a particularly pernicious form of misconduct, denying institutions the opportunity to help students develop subject mastery or educate them on academic integrity.

So how can institutions tackle the threat of AI-to-human rewriter services and ensure students’ work is their own?

1. Increase education on academic integrity. Highlight the value of essential study skills, critical thinking, and originality, and emphasize the importance of the journey, not just the destination.

2. Publicize clear policies around the appropriate use of AI within student workflows. Acknowledge the use of AI as a learning aid but not as a way to bypass learning altogether. Specifically reference AI-to-human rewriter services to avoid doubt.

3. Integrate oral presentations and on-premise examinations to limit opportunities for academic misconduct.

4. Identify students most at risk of misconduct and provide interventions such as writing support, feedback, and mentoring.

5. Be proactive about writing assignments in the age of AI – for example, invite students to share different drafts during the writing process and discuss their thoughts.

6. Implement more advanced AI writing detection software to keep pace with developments in student misconduct and gain a more holistic understanding of students’ writing ability.

7. Look to close opportunities for criminal essay mills to target students via your institutional messaging platforms such as email.

8. Take action on contract cheating. Make students aware of the risk of extortion and blackmail associated with criminal essay mills

Final thoughts: The evolution of essay mills to AI to human rewriter services

Large language models and market demand make students easy prey for criminal enterprises.

Students need to be involved in the discussion around the ethical use of AI – to help them understand the age-old adage that cheating ‘only cheats yourself’ – and to protect them from the risks associated with increasingly predatory and retaliatory actions from essay mills.

Furthermore, a collective response is needed from educators and policymakers to keep pace with the challenges and opportunities of AI in education.

Laws need to be clearer on the acceptable scope of AI use in learning, and institutions must keep pace with how emerging technologies are deployed.