The FCC Wonders how AI will Change the TCPA
Should we regulate up, or regulate down?
FCC Chairwoman Rosenworcel has launched an effort to understand AI’s impact on robocalls and robotexts. Of all the text-messaging initiatives that the FCC is considering, this is perhaps the most far reaching.
The FCC is seeking comments on five questions:
How does AI affect the FCC’s responsibility to enforce the TCPA?
How can we define AI technologies that are relevant to the FCC’s responsibility to enforce the TCPA?
How will AI change the regulatory treatment of the TCPA?
How should the commission differentiate from the deep fakes of voices and texts?
What are the next steps that the FCC should take to understand these issues?
The TCPA is a 1992 statute that is specific to technology that existed at that time—that’s what the Supreme Course said in Facebook v. Duguid. Even post-ruling the TCPA remains a fax-era square peg hammered into a DIY SaaS round hole.
AI will challenge the TCPA’s definitions of type, content, consent, human intervention, and the much-reviled automated telephone dialing system (ATDS). The last one has confounded both plaintiffs and defendants, with both arguing about when human intervention stopped and automated started. Even the Supreme Court called it a fool’s errand:
Duguid contends that ordinary cell phones are not autodialers under his interpretation because they cannot dial phone numbers automatically and instead rely on human intervention. But all devices require some human intervention, whether it takes the form of programming a cell phone to respond automatically to texts received while in “do not disturb” mode or commanding a computer program to produce and dial phone numbers at random. We decline to interpret the TCPA as requiring such a difficult line-drawing exercise around how much automation is too much.
The stakes are high and the FCC is right to ask these questions. It has two options, it can “regulate up”,i.e., start with a blank unregulated slate and let the technology inform how to regulate. The second option is to “regulate down”, in which it applies an existing policy framework and customizes it for AI.
By choosing the first option the FCC risks not responding quickly enough to ensure a business-friendly, consumer-first innovative environment.
The second option can become wasteful and unwieldy to enforce through multiple forbearance rulings in the commission’s effort to bootstrap a law based on 30-year old technology to govern tomorrow’s AI implications.
The real answer, therefore, must come from Congress.
AI is going to change the way we converse, the way we conduct commerce, the way we lead our lives. It requires a head-first dive by the regulator. As Kai-Fu Lee noted, an “AI-friendly policy environment” is one of the four necessary requirements for harnessing the power of AI (abundant data, hungry entrepreneurs, and AI scientists are the other three).
This NOI is a powerful opportunity for the FCC to be visionary and expansive when exploring how AI can protect the consumer. But a true forward-looking legal framework will need to come from Congress if the US is to continue to be a global technology powerhouse.
Thank you for reading and have a great week!
TJ