MW Think robocalls are annoying? AI is making them dangerous.
By Blair Levin and Larry Downes
The FCC is trying to fix customer service by bringing call centers back to the U.S. The real threat: AI-driven scams that regulators can't touch.
AI is lowering the cost of deception while increasing its precision.
Americans are bombarded with robocalls - roughly 4 billion to 5 billion a month - despite decades of regulation.
The Federal Communications Commission's latest proposal to improve customer service aims to regulate offshore call centers. But a bigger shift is already happening. Artificial intelligence is rapidly changing how companies interact with customers - and how scammers target them.
Anyone who has tried to resolve a billing error or cancel a service knows the misery of long waits, impenetrable phone trees and no clear answers. Meanwhile, telemarketing and spam phone calls have reached epidemic levels: Americans are bombarded with robocalls - roughly 4 billion to 5 billion every month - despite decades of regulation.
Congress tried to address the problem with the 1991 Telephone Consumer Protection Act, which included the Federal Trade Commission's "Do Not Call" list. The results speak for themselves. The law was designed for a domestic, voice-based telemarketing industry. Today's robocall ecosystem is global, digital and increasingly automated, allowing telemarketers and scammers to evade detection with ease using technologies the 35-year-old law never anticipated.
That points to a deeper problem with the FCC's proposal. Even if the agency has the authority to require onshoring of call centers, rapid advances in technology make it unlikely such rules would have much impact - even if they extended to online chat and email. For one thing, communications originating abroad - the source of most illegal robocalls - are beyond the practical reach of U.S. regulators.
More fundamentally, the customer-service industry is being scrambled by AI. While the FCC notes AI's potential impact, it doesn't mention that within a few years, most customer interactions by telephone and online will be mediated or entirely handled by better, cheaper and faster AI-based systems.
That shouldn't come as any surprise. Customer-service operations still rely heavily on underpaid and poorly trained humans, leading to general dissatisfaction across all service metrics, including wait times and repeat calls, with consumers increasingly resorting to equally unsatisfactory self-service alternatives.
AI is enabling scalable phishing emails, fraudulent texts, impersonation chatbots, cloned-voice scams and automated outreach.
Meanwhile, AI chatbots have improved at a startling pace in both the accuracy and relevance of their answers to an astonishing range of questions. Though far from perfect, AI systems already outperform human customer-service agents on many routine tasks, particularly in gathering, synthesizing and consistently delivering information. And AI service levels will continue to improve.
But the same technologies that promise better service also create new risks. AI is enabling scalable phishing emails, fraudulent texts, impersonation chatbots, cloned-voice scams and automated outreach across all digital channels. These tools make it easier to target individuals with highly personalized and persuasive messages at scale. As one 2024 study found, AI is lowering the cost of deception while increasing its precision.
The next phase may go further. Agentic-AI systems will act on behalf of users - negotiating, troubleshooting and completing transactions directly with other systems. Instead of calling customer-service lines, consumers may rely on digital agents to resolve problems. While this shift raises new concerns, it also has the potential to eliminate many of the frustrations that plague current systems.
As with many emerging technologies, regulatory approaches that focus on today's market failures may struggle to keep up with rapid change - and risk missing the most important sources of imminent consumer harm. If policymakers want to improve outcomes, they will need to focus on where the problem is going, not where it has been.
Regulators should focus less on prescribing how companies deliver service and more on ensuring transparency and accountability.
That suggests a different set of priorities. Micromanaging the geography of human call-center labor is unlikely to improve outcomes in a world where those workers are increasingly replaced - or bypassed - by software. Regulators should instead embrace tools that address emerging risks, including stronger authentication of communications, improved detection and blocking of fraudulent activity, and clearer disclosure when consumers are interacting with AI systems.
Regulators also should focus less on prescribing how companies deliver service and more on ensuring transparency and accountability. Give consumers better information about service quality and the ability to rate their experiences, creating stronger incentives for improvement.
Then watch how companies respond. Those that fail to adapt will not need to be punished by fines that are often difficult to collect. They will be punished by their customers.
Customer service is changing fast. The real risk isn't that regulators will do too little - it's that they will focus on the wrong problem while a much bigger one grows.
Blair Levin is a nonresident fellow at the Center for Strategic and International Studies and a former chief of staff at the Federal Communications Commission. Larry Downes is co-author of "Big Bang Disruption" and other books on disruptive innovation.
More: Big Tech's AI spending is depriving investors of juicy payouts
Also read: The most prestigious job in AI has no description
-Blair Levin -Larry Downes
This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.
(END) Dow Jones Newswires
May 16, 2026 11:22 ET (15:22 GMT)
Copyright (c) 2026 Dow Jones & Company, Inc.
Comments