A.I. trends in K-12: Fake research, hidden history, and the deepfake policy moment
For Educators · The Social Institute
Credibility is a skill. Here’s what A.I. is doing to it — and how educators can respond.
- A.I. is generating fake academic papers under real professors’ names.
- Researchers are using A.I. to decode centuries-old encrypted documents.
- Boston Public Schools and Paris Hilton are sounding the alarm on deepfakes — and landing on the same conclusion.
When “real” research isn’t real: The new challenge for credibility
A.I.-generated fake citations and fraudulent academic journals are changing what “verified” even means — and that’s a classroom conversation waiting to happen.
The Gist
A.I.-generated academic papers are using real professors’ names without their knowledge and are getting published in fake journals that look legitimate. Plus, researchers are now seeing a surge of “phantom citations”: references to studies, books, and authors that simply don’t exist. As A.I. gets more embedded in education and research, teaching students to verify sources isn’t optional anymore.
What to Know
NBC News highlights how A.I. is creating realistic-looking citations that can fool reviewers, researchers, and even the scholars whose names are attached to them. In one example, two education professors were asked to share a paper they supposedly published in 2023, except the paper never existed. The citation included a journal name, volume number, page range, and even a fake DOI link. Journal editors say these A.I.-generated references are becoming increasingly difficult to detect because they often use real authors, real journals, and believable publication details.
The challenge extends beyond fake citations. A network of fraudulent academic journals has published more than 100 A.I.-generated papers, sometimes using the names of real professors without their knowledge. The output often looks authoritative even when it’s completely made up. With A.I. able to generate hundreds of polished references in minutes, a professionally formatted source is no longer proof of a real one.
TSI’s Take
A.I. can already copy photos or personal information. It can imitate someone’s voice, writing style, expertise, and online identity. In a world where A.I. can generate convincing content in seconds, credibility becomes a superpower.
Here are three ways educators can help students navigate this new reality:
- Teach students to verify: Before sharing information, encourage students to double-check claims using trusted sources or fact-checking websites.
- Create a culture of responsible content creation: Help students understand that A.I.-generated content should be reviewed, edited, and fact-checked before it’s shared.
- Encourage students to speak up when something seems off: If students encounter questionable information online, they can check with a teacher, compare multiple sources, or flag the content for further review.
Can your students tell the difference between what’s real online and what’s not? Do they have the skills to do so? The #WinAtSocial Lesson, Fake A.I. content, helps students learn how to spot A.I.-generated images and videos, think critically about online content, and make responsible choices when something doesn’t seem right.
Unlocking centuries of hidden history with A.I.: Bringing the past into today’s teaching
What if A.I. could uncover centuries-old secrets hidden in plain sight? For historians, that’s no longer hypothetical — and for educators, it’s a powerful way to show students what A.I. can do.
The Gist
What if A.I. could uncover centuries-old secrets hidden in plain sight? An article by BBC explores how researchers are using artificial intelligence to decode encrypted historical documents that have remained unread for hundreds of years, revealing everything from love letters and political plots to medical remedies. By combining machine learning, handwriting recognition, and language models, historians are gaining access to information that could reshape our understanding of the past. The story highlights A.I.’s power as a tool for discovery, pattern recognition, and making previously inaccessible knowledge available to learners.
What to Know
Researchers are using A.I. to crack historical ciphers found in archives and libraries around the world. According to estimates, roughly 1% of all archival materials are fully or partially encrypted, meaning a significant amount of history remains hidden. A.I. is helping historians transcribe difficult handwriting, recognize unusual symbols, and identify patterns in coded texts much faster than traditional methods. Beáta Megyesi, a professor in computational linguistics at Stockholm University, described the work as uncovering “someone’s secrets and a lost historical world,” emphasizing how these documents can fill gaps in historical narratives and even change what we know about major events and figures.
The technology is already producing impressive results. Researchers developed an A.I. chatbot capable of both transcribing and decoding historical ciphers, and when tested on the Borg Cipher, it successfully translated a 500-symbol passage in just 29 minutes while also explaining its reasoning. Beyond speed, A.I. is helping historians analyze documents at a scale that would be impossible manually, opening doors to discoveries hidden in centuries of letters, records, and manuscripts. Suddenly, documents long locked away can now be read and used, reshaping the stories and sources students encounter in the classroom.
TSI’s Take
This story offers a powerful example of how A.I. can expand human knowledge rather than simply automate tasks. Moreover, the most impactful uses of A.I. often happen when technology and human expertise work together.
Educators can use this story as a starting point to:
- Show students that A.I. can be a tool for discovery: A.I. can help uncover hidden knowledge, preserve history, and solve complex problems that humans have been working on for centuries.
- Build students’ critical thinking skills around A.I.: As A.I. tools become more capable, students need practice questioning results, verifying information, and understanding that A.I. outputs still require human oversight and expertise.
- Emphasize the human role in the process: Students should understand that technology works best when paired with human curiosity, judgment, and ethical decision-making.
Hidden history only becomes useful when we learn how to ask the right questions. The #WinAtSocial Lesson, A.I. that works for us, shows how clearer prompts lead to better thinking, helping students use A.I. in a more intentional, thoughtful way.
Boston Public Schools and Paris Hilton are sounding the alarm on deepfakes
Two very different stories, one shared message: teaching students to think about A.I. is just as important as teaching them how to navigate it.
The Gist
A.I. isn’t waiting for schools to catch up, and Boston Public Schools is taking a proactive approach, and it’s not just about plagiarism. The district is drawing a hard line on deepfakes, privacy, and what responsible A.I. use looks like in schools. Meanwhile, Paris Hilton is partnering with a journalist on a TikTok docuseries investigating A.I.-generated deepfakes and digital impersonation. Two very different stories, one shared message: teaching students to think about A.I. is just as important as teaching them how to navigate it.
What to Know
Boston Public Schools’ proposed policy covers everything from student privacy to the growing concern of A.I.-generated deepfakes. Students and staff would be prohibited from creating fake images, videos, or audio of real people without consent, and any harmful or misleading content would need to be reported and addressed by school officials. The policy also makes it clear that A.I. should never be the sole factor in grading, discipline, or academic evaluations.
The district’s focus on deepfakes reflects a growing challenge far beyond the classroom. Recently, Paris Hilton partnered with journalist Laurie Segall on a TikTok docuseries investigating the impact of A.I.-generated deepfakes and digital impersonation. The project highlights how rapidly advancing technology is creating new concerns around consent, privacy, misinformation, and online safety. Digital citizenship isn’t just about social media manners anymore. It’s about understanding how A.I. can be used responsibly, and what it looks like when it isn’t.
TSI’s Take
Boston Public Schools’ policy shows how schools are now thinking about A.I. in two directions at once: teaching it as a literacy and setting guardrails around its misuse. The #WinAtSocial Lesson, School tech policies: Devices, A.I., and beyond, was built for exactly this conversation.
Educators can help students understand the reasoning behind A.I. by:
- Connecting policy to real-world examples: Use current events, such as the rise of deepfakes and digital impersonation, to help students understand why schools are creating A.I. guidelines in the first place.
- Teach responsible decision-making: Encourage students to think critically about when A.I. is an appropriate tool, when it isn’t, and how its use can impact others.
- Make A.I. literacy part of digital citizenship: Discussions about privacy, consent, misinformation, and online responsibility are becoming essential skills for navigating an increasingly A.I.-driven world.
How can you protect students when cyberbullying now includes A.I.-generated deepfakes? The Social Institute’s School Playbook, Administrators’ guide to empowering students to tackle cyberbullying in the age of A.I. deepfakes, gives administrators a practical, research-backed guide to help students recognize, respond to, and prevent harm in today’s A.I.-driven digital world.
Students need practice slowing down and questioning what they see, and teachers are now the ones best positioned to build that habit. The Social Institute makes it easy to start. Explore ready-to-use lessons, download the playbook, or request a demo to see how #WinAtSocial brings these conversations into the classroom.