Criminalizing digital impersonation on an indistinguishable-from-real standard — including invented people
S.3982, the AI Fraud Accountability Act of 2026, amends section 223 of the Communications Act of 1934 to make it unlawful to use digital impersonations to commit fraud in interstate or foreign communications. Digital impersonation is defined by whether a reasonable person, viewing or listening as a whole, could distinguish it from an authentic depiction.
Bill overview (primary data)
- Bill numberS. 3982
- TypeSenate Bill
- Congress119th Congress
- Latest actionRead twice and referred to the Committee on Commerce, Science, and Transportation.（2026-03-04）
Key points
- S.3982 amends section 223 of the Communications Act of 1934 to criminalize using digital impersonations to commit fraud.
- The test is whether a reasonable person viewing or listening as a whole could distinguish it from the real thing — how it is received, not how it was made.
- It covers both impersonation of real individuals and depictions of imaginary individuals indistinguishable from real ones.
- Identifiable individual reaches beyond face, likeness and voice to a unique birthmark or other recognizable feature.
- Rather than new law, it connects to an existing framework for crimes committed by means of communication.
1Setting the test at whether it can be told apart
Where to begin regulating generated video and audio depends on the state of the technology. Since quality keeps improving, a line drawn at technical specification goes stale quickly. The test this bill adopts is whether a reasonable person, viewing or listening as a whole, could distinguish it from the real thing. Judgment turns on how it appears to the recipient rather than on how it was made.
2Two categories
The second category shows the care in the definition. Where a person who does not exist has been fabricated, there is no impersonated victim. Yet if the result is indistinguishable from a depiction of a real person, the recipient believes they are seeing a human being. Because the effect as a fraud technique is the same, a category is provided that does not require identifying a victim.
3Defining identifiable individual
The text reaches beyond face, likeness and voice to a unique birthmark or other recognizable feature. Even with no face shown and no voice used, a distinguishing feature that makes someone recognizable can bring a depiction within scope — a definition written to survive technology moving toward partial reproduction.
4Choosing to amend the Communications Act
What is amended is section 223 of the Communications Act of 1934, a provision that has dealt with harassment and threats made through communications. Fraud by digital impersonation is added there. Rather than writing new law, the bill connects to an existing framework for crimes committed by means of communication.
Another bill this site covers writes AI into the campaign law prohibition on misrepresentation, so the pattern of inscribing AI into existing prohibitions recurs. Of the 120 bills this site holds as of 2026-09-02, 94 (78 percent) remain referred to committee.
Why it matters
For businesses handling generated media, regulatory reach may extend past reproducing real people to fabricating people who seem real. A test based on reception is durable against technical progress but makes ex ante legality assessment harder.
FAQ
Why not draw the line at technical specification?
Why cover imaginary people?
Sources (primary)
Source: Congress.gov (Library of Congress; U.S. legislative materials, public domain). Links go to the official site.
- Congress.gov (bill page, original)
- S. 3982（119th Congress）