CoSN: Cybersecurity and Data Privacy Remain Top AI Concerns in Education

A leading concern for education technology leaders across the United States is the potential for AI to enable new forms of cyber attacks, according to the latest State of Ed Tech report from CoSN. The professional association for K-12 ed tech leaders surveyed 607 of its constituents across the country to find out how they are navigating the ever-changing technology landscape, including issues such as artificial intelligence, cybersecurity, connectivity, staffing, procurement, and device management.

Seventy-five percent of survey respondents cited AI-enabled cyber attacks as very concerning, while another 23% said they were moderately concerning. Other primary concerns include:

  • Student data privacy (62% of respondents said they were "very concerned");
  • Lack of teacher training for integrating AI into instruction (52%);
  • Spread of false information (50%); and
  • New forms of cyberbullying (50%).

"These concerns underscore the importance of clear policies, safeguards, and professional learning as AI use expands," the report noted.

Perhaps equally important were the areas that few ed tech leaders were very concerned about:

  • AI surpassing humans (only cited by 9% of respondents);
  • Overall job loss (8%); and
  • The technology will replace teachers (5%).

Generally, ed tech leaders were optimistic about AI's potential to positively impact education. Respondents saw productivity as the area where AI has the greatest potential (cited by 74%), followed by personalized education (67%), student tutoring (46%), and preparing students for the workforce (43%).

About 88% of respondents reported that their districts have initiative in place to support AI use. The most common AI initiative was training instructional staff on the use of instruction-focused generative AI tools, cited by 70% of respondents. Other common initiatives included:

  • Productivity suite platforms for administrative/support staff (54%);
  • Productivity suite platforms for teachers/instructional staff (53%);
  • Training for administrative/support staff on the use of productivity suite platform tools (51%); and
  • Instructional platforms for teaching and learning use (41%).

In light of ed tech leaders' cybersecurity and data privacy concerns, it's interesting to note that just 7% of districts reported engaging in initiatives to prepare identity access management systems or prepare data storage environments for the age of AI.

The full report is available on the CoSN site here.

About the Author

Rhea Kelly is editor in chief for Campus Technology, THE Journal, and Spaces4Learning. She can be reached at [email protected].

Featured

  • abstract agentic ai data streams

    Rubrik Brings Agentic AI to Its Flagship Cyber Resilience Platform

    Rubrik has introduced Rubrik AI, positioning the company's cyber resilience platform around agentic operations across data, identity, and AI infrastructure.

  • Futuristic glowing digital book

    AI Won't Replace Teachers — But It May Be What Makes Structured Literacy Work at Scale

    Purpose-built AI systems can analyze patterns in student performance, identify specific skill gaps, and connect those gaps directly to instructional recommendations. Done well, this doesn't remove the teacher from the equation. It sharpens the teacher's ability to act.

  • tool icons with variety of business icons

    SETDA Releases Free EdTech Quality Action Toolkit

    The State Educational Technology Directors Association (SETDA) has put together a free K-12 EdTech Quality Action Toolkit that provides a framework for evaluating education technology products as well as guidance on regulatory compliance, templates for communicating with vendors, training resources, and more.

  • circuit patterns

    Anthropic Intros Lower-Cost Claude Sonnet 5

    Anthropic has launched Claude Sonnet 5, positioning the model as its most autonomous mid-tier offering to date and a lower-cost alternative to its flagship Opus 4.8 system. The company said the model can plan multi-step tasks, operate tools such as browsers and terminals, and complete agentic work at a level that previously required larger and more expensive models.