Deepfakes – National Center for State Courts

“The Fire and BPA’s Preservation of Evidence”
March 4, 2026
Book Review: Craig Ball, “Forensic Tells: The Litigator’s Guide to Detecting Deepfakes and Authenticating Digital Evidence”
March 7, 2026

Evaluating deepfakes, the National Center for State Courts asked: “Is anything ever truly new again?”  “Evidentiary Issues Raised by Artificial Intelligence: A Policy Paper” (Apr. 2025).

In answering its own question—“What can courts do?”–the Center suggests the following questions that a court can pose to a proffering party to help determine whether a deepfake is being presented:

  1. What is the source of this digital evidence, and when, how, and where was it obtained?
  2. Who has had access to this evidence from the time it was created or captured until now?
  3. What steps were taken to preserve the original media or data and prevent tampering or alteration?
  4. Can you tell me who has had custody of this evidence, including any sharing or , transfer of the evidence, and the places it has been stored (chain of custody)?
  5. Has the evidence been altered, edited, or processed in any way since it was originally created?
  6. Is there any other data or source that can help confirm its authenticity? Specifically, if the proffered digital evidence is a text message thread that is challenged on authenticity grounds by the opposing party, the court should request that the challenging party submit the corresponding text message thread on their device for verification purposes.

The Center adds “Even in cases where parties are self-represented, these questions can be appropriate. The next set of questions likely requires more expertise than many litigants may have.”

  1. Were any forensic tools or methods used to verify the integrity of the evidence?
  2. Can you provide metadata or other information that supports the authenticity of this digital file?
  3. Can a qualified expert explain the processes used to handle and verify this digital evidence?

The Center also suggests consideration of the type of case as a factor in authenticating proffered evidence:

Analysis of digital evidence by the court must take into consideration the type of case and the discipline involved. For example, AI-generated evidence in a family law case, such as a demonstrative video submitted by a spouse in a child-custody case simulating an alleged assault inflicted upon them by the other spouse, should be regarded with appropriate scrutiny when determining its admissibility, as it may not have been offered for its probative value. Similarly, in cases involving restraining orders or civil harassment, text messages or voice recordings purportedly from the accused party must be carefully authenticated, given the ease with which AI can now mimic voices and writing styles. School violence cases might involve fabricated social media posts or manipulated surveillance footage, requiring thorough verification. The admissibility of such evidence can be further complicated by hearsay rules, especially when dealing with AI-generated text messages or emails.

Addressing the burden of proof, the Center asks: “Should the party introducing digital evidence be required to provide comprehensive authentication documentation? Should the opposing party shoulder the burden of proving manipulation? Or, when one or both parties are self-represented litigants, does the court have any responsibility to support or challenge the proffered digital evidence?”

In answering those questions, it wrote that:  “The technical complexity of deepfake detection may necessitate new standards for expert testimony and verification procedures. As verification and detection products improve and become both more affordable and more widely available, courts will need to continually assess the extent to which they should be actively involved in evaluating digital evidence.”

At bottom, the Center notes that these are “unresolved questions” that require “continued dialog….” It states: “For now, the solution lies not in entirely reinventing our approach to evidence but in thoughtfully applying and adapting existing principles to new technological contexts.”

The Center provides two resources:

Share