Legal scholar Cass R. Sunstein wisely notes that America’s freedom of speech never has been absolute. Perjury, false advertising and other types of untruths are illegal. He argues that courts can and should remove First Amendment protection for harmful falsehoods, such as so-called deepfake videos and images of public figures that liars create by harnessing artificial intelligence. Misinformation has become ubiquitous as false statements proliferate online, particularly on Facebook, Twitter and other social media platforms. Widespread falsehoods about government officials and agencies generate negative consequences, Sunstein writes, and endanger democratic self-government.
The United States should limit free speech by punishing people who spread harmful falsehoods.
US Supreme Court justice Oliver Wendell Holmes Jr. made the famous statement that free speech does not permit someone to falsely yell “fire” in a crowded theater. Yet today, many people make false claims equivalent to yelling “fire” in a crowded theater, bringing consequences ranging from illness and death to the destruction of democratic self-government.
“Falsehoods” include both false statements by people who think they’re true and lies by people who know they’re lying. These distinctions matter in determining what, if any, regulatory penalty should apply to a liar or to someone who makes a mistake. Penalties for falsehoods could discourage robust dissent. Generally, self-governed societies protect liars. Correcting falsehoods is usually better for democracy than censoring or sanctioning them might prove to be.
The US Constitution allows the government to do more to deter defamation – and it should, by punishing falsehoods that threaten public health, safety and the functions of democracy.
However, banning all falsehoods would “crush...
Cass R. Sunstein, the Robert Wesley University professor at Harvard University, founded and directs the Program on Behavioral Economics and Public Policy at Harvard Law School. His other books include How Change Happens and Too Much Information.