Senate-Passed Bill Outlawing AI-Generated Explicit Deepfakes Moves To House
The U.S. Senate on Thursday unanimously approved the Disrupt Explicit Forged Images and Non-Consensual Edits Act, or DEFIANCE Act, a bipartisan effort to strengthen legal protections for individuals targeted by nonconsensual deepfake imagery. The bill now proceeds to the House of Representatives for consideration.
The measure would allow individuals depicted in sexually explicit deepfake images or videos — digital fabrications created without their consent — to pursue civil damages of at least $150,000 per violation against persons responsible for creating or sharing such content.
Legislators who supported the bill said existing legal frameworks are insufficient to address the growing prevalence of deepfake technology and the unique harms it causes. They noted the legislation builds on earlier federal and state laws aimed at curbing nonconsensual intimate imagery, but expands the scope and clarity of remedies available under federal civil law.
Deepfake content — synthetic media produced using artificial intelligence and machine learning — has surged in recent years, raising concerns in Washington about privacy, harassment, fraud, and national security. Lawmakers from both parties have pushed a series of proposals in recent sessions to update laws governing digital impersonation and nonconsensual imagery.
Earlier legislative efforts focused on criminal penalties for creating or distributing explicit deepfakes of public officials or election candidates, or unauthorized alterations of videos used in political context. Other bills aimed to enhance law enforcement’s ability to investigate and prosecute deepfake-related fraud and identity theft.
The DEFIANCE Act differs from those proposals by creating a federal civil right of action, enabling private individuals — not just government prosecutors — to seek monetary damages in federal court. The bill would supplement state laws that vary widely in enforcement and penalties related to deepfake and revenge-porn imagery.
Supporters have argued that civil remedies are crucial because many victims face ongoing reputational harm and emotional distress long after illicit content is published. Civil suits, proponents say, can provide both compensation and deterrence.
If the House approves the DEFIANCE Act and the president signs it into law, the new provisions would expand legal avenues for victims of nonconsensual deepfakes and related digital forgeries. Advocates for stronger protections have said the approach could serve as a model for future legislation addressing other forms of digitally manipulated content.
Opposition in the Senate was nonexistent, reflecting bipartisan agreement on the need to update legal tools in the face of rapid advances in artificial intelligence and digital media technologies.
The bill’s proponents say it represents a significant step in the federal government’s response to technology that can create convincing but fraudulent depictions of real people, often used to harass, humiliate or exploit victims.
Meanwhile, social media influencer and entrepreneur Paris Hilton joined Democratic Rep. Alexandria Ocasio-Cortez in announcing a new collaborative effort this week aimed at combating the creation and distribution of AI-generated sexually explicit imagery without consent.
The initiative, unveiled Thursday, seeks to raise awareness of the growing prevalence of artificial intelligence tools that can produce realistic deepfake pornography using the likenesses of real individuals. The effort calls for legislative and technological solutions to protect potential victims and hold creators and distributors accountable.
Hilton, who has previously spoken publicly about being targeted by nonconsensual explicit content earlier in her career, said that the proliferation of AI tools “makes it easier than ever” for deceptive imagery to spread and cause harm. She urged lawmakers and technology companies to act urgently to establish safeguards.
Ocasio-Cortez, a member of the House Committee on Energy and Commerce, emphasized the need for stronger legal frameworks that can address the unique challenges posed by AI-generated content. She noted that traditional privacy and harassment statutes may not fully encompass the nuances of AI-enabled manipulation.
The pair’s announcement follows increased attention from lawmakers and advocacy groups concerned about how advances in artificial intelligence intersect with issues of consent, privacy and online safety. AI researchers and civil liberties organizations have also called for clearer standards and potential regulatory measures to limit the misuse of image synthesis technologies.
Nearly 150,000 California Mail Ballots Rejected In Primary
California election officials are examining why nearly 150,000 mail-in ballots were rejected during the state’s June primary, with late-arriving ballots accounting for most of the increase.
The higher rejection rate comes despite years of election law changes designed to make voting more accessible and ensure more ballots are ultimately counted.
According to data compiled by the California Secretary of State’s office, 148,241 mail-in ballots were rejected during the June primary, representing 1.73% of all mail ballots returned.
That marks an increase from the 2024 primary, when 108,982 ballots were rejected, or 1.56% of all mail ballots cast.
The largest reason for rejection was ballots arriving too late to qualify under California law.
State data shows that 93,479 ballots were rejected because they failed to meet the state’s deadline requirements.
California allows mail ballots to arrive up to seven days after Election Day, provided they are postmarked on or before Election Day.
Voting experts believe many of the rejected ballots received postmarks dated after Election Day, making them ineligible to be counted even if they arrived within the seven-day grace period.
Kim Alexander, president of the nonpartisan California Voter Foundation, said the timing of postal processing appears to be a significant concern.
“Ballots rejected due to lateness are caused primarily to being postmarked too late to count, not because they arrived too late to count,” Alexander said.
Election experts have questioned whether recent operational changes within the U.S. Postal Service may be contributing to delayed postmarks.
Earlier this year, a group of mostly Democratic U.S. senators wrote to the Postal Service expressing concern that mail-processing changes could affect ballot handling during federal elections.
Updated Postal Service procedures mean postmarks may reflect when mail is processed at regional facilities rather than when it first enters the mail system.
Because processing centers have been consolidated in some areas, election officials have warned that postmark timing could be affected.
Before the June primary, California election officials encouraged voters to mail ballots as early as possible because of potential postal delays.
The June election also featured a highly competitive race for governor, which some experts believe encouraged many voters to wait until the final days before casting ballots.
In addition to late-arriving ballots, approximately 44,000 ballots were rejected because election officials determined the voter’s signature did not match the signature on file.
Another 8,300 ballots were rejected because the return envelope did not contain a signature.
State records also show that 743 ballots were rejected because those voters had already cast another ballot.
Other rejected ballots involved missing ballots inside return envelopes or multiple ballots submitted in a single envelope.
The highest rejection rate in the state occurred in Tulare County, where 3.52% of returned mail ballots were rejected.
Alpine County and Merced County followed closely, each recording rejection rates of 3.36%.
California has frequently faced criticism for taking weeks to complete vote counting after major elections, largely because state law permits ballots meeting certain requirements to continue arriving after Election Day.
Following the June primary, President Donald Trump again criticized California’s election system, while the Department of Justice opened an investigation into election administration in Los Angeles County.
Election experts emphasize that the increase in rejected ballots should not be interpreted as evidence of widespread voter fraud.
A 2025 report by the Brookings Institution found documented cases of mail ballot fraud remained extremely rare, estimating roughly four cases for every 10 million mail ballots cast, The Associated Press reported.
State officials have not identified evidence of widespread fraud connected to the June primary, but the increase in rejected ballots is expected to receive additional scrutiny as election administrators evaluate whether changes in postal operations, voter behavior or election procedures contributed to the higher rejection rate.
