Since 1776, the American commitment to free speech has always been remarkable. The First Amendment does not only protect popular ideas–it protects offensiveness, anonymity, and dissent. It reaffirms that the best response to bad speech is more speech, not enforced silence.
Today, we are facing a technological shift. As AI lowers the cost of producing and distributing highly realistic content, it is no longer just adding more speech to the marketplace of ideas. In some instances, it is being used to hijack individuals’ identities to deceive the public. Today, anyone with the right tools can generate a video of a person saying something they never said in a voice indistinguishable from their own.In recent election cycles, AI-generated robocalls impersonating public officials–such as a fake Joe Biden call in New Hampshire–and the growing use of deepfake media have raised significant concerns about their potential to mislead or suppress voters (PBS NewsHour; FCC). While this threat is often framed as a problem of “misinformation” to justify broad new censorship powers, it is more accurately understood as a challenge to attribution and individual agency.
Deepfakes do not merely produce content: they can be used to reassign identity by attaching words and actions to individuals who never expressed them.
The First Amendment, at its core, protects acts of human expression. Even when courts extend those protections broadly, they still assume that speech reflects some form of intent, but AI disrupts that assumption. Deepfakes do not merely produce content: they can be used to reassign identity by attaching words and actions to individuals who never expressed them.
However, in addressing these precise simulations, we must ensure that the cure is not worse than the disease. Much of the current debate surrounds whether AI-generated content should be protected under the First Amendment. Organizations like the Foundation for Individual Rights and Expression (FIRE) have rightly emphasized that AI, like the printing press or the internet, should not become an excuse for broad speech restrictions. Any regulation must remain narrowly tailored to avoid creating a “censorship office” within the government. A recent FIRE poll found that 92% of Americans want the government to protect free speech, even as they consider AI regulation.
What is needed is a coherent application of existing principles. We must recognize that the problem is not that AI speech might be protected; it is that our current framework assumes speech belongs to a speaker. When that assumption is exploited through deceptive impersonation, it is the individual’s right to speak–and their right not to be used as a puppet–that is under fire.
To address this gap without compromising the First Amendment, the law should look to existing traditions that recognize limits where speech overlaps with the rights of others. Defamation law protects individuals from false statements that harm reputation. Fraud law prohibits deceptive misrepresentations for gain. These are narrow, historically grounded exceptions that address false and harmful speech case-by-case, without empowering the government to define truth on a broad scale.
The First Amendment protects anonymous speakers–not fabricated ones.
Any federal statutory framework must be designed with extreme precision to avoid becoming a tool for repression. It must include ironclad protections for parody, satire, and social commentary. A “right to self-expression” must include the right to create transformative works, even those that use AI. Therefore, any legal remedy must be limited to realistic, deceptive uses of identity that lack a clear expressive or transformative purpose.
By focusing on consent and deception rather than the truth or falsity of the content, this approach targets whether a speaker is being misrepresented, not whether their ideas are correct. In other words, it protects the authenticity of the speaker’s identity rather than regulating the substance of what is said. This distinction is crucial for maintaining the First Amendment’s commitment to viewpoint neutrality, because the government is not evaluating competing ideas but ensuring that speech is not falsely attributed.
Deepfakes crystallize the need for this clarity. They do not merely say something about you–they attempt to be you. Recognizing the importance of individual attribution does not weaken the First Amendment; it reinforces it. Free speech depends on the ability to evaluate credibility and engage in dialogue grounded in human agency. If identity can be freely appropriated by the state or bad actors to manufacture “consent” or “speech” that never happened, the foundation of the First Amendment erodes. The First Amendment protects anonymous speakers–not fabricated ones.
A free society requires that our voices remain tied to our real identities.
Constitution doctrine has evolved before in response to shifts in how power operates. The First Amendment has always stood for the freedom of the individual. In the age of AI, that commitment demands something more: not the right to be protected from “false” ideas, but the right to ensure that your own voice remains your own. The issue is no longer just who is allowed to speak, but ensuring that when a person speaks, the world knows it is actually them. A free society requires that our voices remain tied to our real identities, ensuring that the freedom promised in 1776 survives the digital age.
Alandee Patil ‘29 studies Political Science and PNP in the College of Arts & Sciences. She can be reached at apatil@wustl.edu.