K12 and California Virtual Academies Settle with State

Attorney General Harris announces $168.5 million settlement; K12 says it’s only $8.5 million.

A for-profit company that operates online charter schools in California has reached a $168.5 million settlement with the state’s attorney general's office over allegations that it manipulated attendance records, engaged in false advertising and overstated the academic progress of students.

Attorney General Kamala Harris announced Friday that Virginia-based K12 Inc. will pay $8.5 million to settle the state’s claims. The agreement is still subject to court approval.

Harris said K12 also agreed to provide $160 million in debt relief to the nonprofit schools it manages — which in California are known as the California Virtual Academies (CAVA).

As part of the settlement, K12 will also adopt about 60 corrective actions, including undergoing independent reviews of its services for students with disabilities, providing teachers with sufficient information and training to prevent improper claiming of attendance dollars, ensuring the accuracy of all advertisements and changing policies and practices to prevent the kinds of conduct that led to the attorney general’s investigation.

"All children deserve, and are entitled under the law, to an equal education," Harris said in a prepared statement. "K12 and its schools misled parents and the state of California by claiming taxpayer dollars for questionable student attendance, misstating student success and parent satisfaction and loading nonprofit charities with debt."

K12 said in a statement that it admitted no wrongdoing in the settlement and it had already planned to adopt many of the corrective actions required under the agreement. K12 also disputed the attorney general’s description of the amount of the settlement, calling it "flat wrong."

"Despite our full cooperation throughout the process, the Office of the Attorney General grossly mischaracterized the value of the settlement just as it did with regard to the issues it investigated," K12 CEO Stuart Udell said in a prepared statement.

Udell said the debt relief, or credits, should be called "subsidies," not debts, and that the company’s agreement to expunge them shouldn’t be used by the attorney general to hike the size of the settlement. He also defended the credits, saying they had protected the schools against financial uncertainties, given the state’s funding environment.

An investigation by the East Bay Times and the Bay Area News Group revealed that K12 reaps tens of millions of dollars annually in state funding while graduating fewer than half of its high school students. It also showed that children who spend as little as one minute per school day logged onto K12’s software may have been counted as "present" in records used to calculate the amount of funding the schools get from the state.

In its investigation, the attorney general’s office alleged that K12 and the 14 CAVA schools it operates in California misled parents to induce them to enroll their children in K12-run schools by publishing misleading advertisements about students’ academic progress, parent satisfaction, class sizes, hidden costs, the quality of the materials provided to students and their graduates’ eligibility for admission into the University of California and California State University systems.

CAVA and K12 are currently undergoing an audit by the state controller’s office and the California Department of Education (CDE). The audit will continue despite the settlement with the attorney general’s office, said Robert Oakes, assistant director of communications for the CDE.

"These are on parallel but separate tracks," Oakes said. "The AG’s (office) is doing their own thing. It’s not connected. The audit stays underway."

Approximately 15,000 students are enrolled in CAVA’s virtual charter schools. A recent study found that students in virtual schools that exist solely online are far behind their peers in math and reading. Furthermore, reports indicate that CAVA schools collectively had a graduation rate of 36 percent, compared to the state average of 78 percent.

About the Author

Richard Chang is associate editor of THE Journal. He can be reached at [email protected].

Featured

  • stylized illustration of a desktop, laptop, tablet, and smartphone all displaying an orange AI icon

    Survey: AI Shifting from Cloud to PCs

    A recent Intel-commissioned report identifies a significant shift in AI adoption, moving away from the cloud and closer to the user. Businesses are increasingly turning to the specialized hardware of AI PCs, the survey found, recognizing their potential not just for productivity gains, but for revolutionizing IT efficiency, fortifying data security, and delivering a compelling return on investment by bringing AI capabilities directly to the edge.

  • handshake between two individuals with AI icons (brain, chip, network, robot) in the background

    Microsoft, Amazon Announce New Commitments in Support of Presidential AI Challenge

    At the Sept. 4 meeting of the White House Task Force on Artificial Intelligence Education, Microsoft and Amazon announced new commitments to expanding AI education and skills training.

  • digital learning resources including a document, video tutorial, quiz checklist, pie chart, and AI cloud icon

    Quizizz Rebrands as Wayground, Announces New AI Features

    Learning platform Quizizz has become Wayground, in a rebranding meant to reflect "the platform's evolution from a quiz tool into a more versatile supplemental learning platform that's supported by AI," according to a news announcement.

  • abstract pattern of cybersecurity, ai and cloud imagery

    Report Identifies Malicious Use of AI in Cloud-Based Cyber Threats

    A recent report from OpenAI identifies the misuse of artificial intelligence in cybercrime, social engineering, and influence operations, particularly those targeting or operating through cloud infrastructure. In "Disrupting Malicious Uses of AI: June 2025," the company outlines how threat actors are weaponizing large language models for malicious ends — and how OpenAI is pushing back.