New California Laws Aim to Prevent AI Harm during Elections


    25 September, 2024

    California recently enacted three legislative measures that ban the use of artificial intelligence (AI) tools to create false images and videos during election campaigns.

    The laws – which are among the strongest in the U.S. – aim to reduce the use of AI to produce fake materials that could mislead voters during elections. California Governor Gavin Newsom signed the bills into law last week.

    One of the main goals of the California measures is to fight the use of so-called deepfakes. A deepfake is a piece of video or audio created to make it appear that people in it are saying or doing things that they never said. Deepfakes have already been used in political campaigns around the world.

    In this file photo, California Gov. Gavin Newsom speaks at the opening ceremony for Panda Ridge, the new exhibit at the San Diego Zoo, Aug. 8, 2024, in San Diego. (AP Photo/Derrick Tuskan, File)
    In this file photo, California Gov. Gavin Newsom speaks at the opening ceremony for Panda Ridge, the new exhibit at the San Diego Zoo, Aug. 8, 2024, in San Diego. (AP Photo/Derrick Tuskan, File)

    One of the laws requires large online publishers to either remove or identify content that has been digitally created or changed in an effort to mislead voters during specific election periods.

    Another requires any election advertisements that have been created or changed using AI methods to include a message informing the public that the material has been changed.

    The third bill bans the publishing of certain kinds of misleading information about campaign activities within 120 days of an election.

    All the California measures permit state law enforcement officials to take different forms of action against violators of the laws.

    In a statement issued after he signed the bills, Newsom said the measures were aimed at “safeguarding the integrity of elections.” He added, “It's critical that we ensure AI is not deployed to undermine the public's trust through disinformation – especially in today's fraught political climate.”

    State lawmakers in several other U.S. states have passed similar measures in recent years to prevent deepfakes and misinformation during election campaigns. But critics argue that the laws are difficult to enforce and face numerous legal actions aimed at overturning them.

    The Associated Press (AP) reports two of the new California measures immediately faced legal action seeking to block the legislation. One of the lawsuits argues that one of the measures censors expressions of free speech and permits anyone to take legal action against material they do not agree with.

    That lawsuit was brought by an individual who created parody videos that received public attention online. This included changed audio of Vice President and Democratic presidential nominee Kamala Harris. At least one of the creator's videos was shared by Elon Musk, who owns the social media service X.

    The governor's office told the AP the law does not ban satire and parody content. Instead, it requires material changed through AI methods to be clearly identified in videos, audio or images. A spokesman for Newsom, Izzy Gardon, said in a statement, “It's unclear why this conservative activist is suing California.” Gardon added that California's measure is nearly the same as laws passed in other states.

    But critics – such as free speech activists and Musk – have called the new laws unconstitutional. They have argued they violate the U.S. Constitution's First Amendment, which protects speech and other personal freedoms.

    Ilana Beller is with Public Citizen, a nonprofit group that seeks to protect individual rights. She told the AP she does not know how effective such laws will be in stopping election deepfakes. Beller noted her group closely follows U.S. measures related to deepfakes and none of them have yet been tested in court.

    Some technology experts warn that the laws' effectiveness could be limited by the slow process of court actions. Misleading content, including deepfakes, might not be removed before harm has been done, the experts say. Even if it only takes a court a few days to act against such material, candidates and the election process may have already been hurt, Beller said.

    I'm Bryan Lynn.

    Bryan Lynn wrote this story for VOA Learning English, based on reports from The Associated Press, Reuters and the California Governor's office.

    ______________________________________

    Words in This Story

    fake – adj. false or not real

    integrity – n. the quality of being honest and having strong moral principles that you refuse to change

    undermine – v. to make someone less confident of make something weaker

    fraught – adj. full of danger or difficulties

    censor – v. to examine books, documents or films in an attempt to remove part that are offensive or not permitted by rules

    parody – n. a kind of content that copies someone's else's style in a funny way

    satire – n. the use of jokes and humor to criticize people or ideas