AI Calling: “Press 1 to Cancel” Scams Cross Borders as AI Makes Them Cheaper and More Personal
Published
Key Takeaways
- AI makes it difficult to distinguish malicious communications from normal ones due to the volume of realistic messages.
- Quilici warns that AI scams leverage data from data brokers and breaches, allowing calls feel credible.
- YouMail identifies scams within a few calls because they follow similar patterns with small changes in the product, price, and brand details.
- Scammers adapt faster than enforcement mechanisms and function across borders to evade carrier controls.
- In 2025, robocalls held steady at 52.5 billion, but unwanted calls surged more than 15% year over year.
We spoke with Alex Quilici, CEO at YouMail, about how robocalls and scam campaigns are evolving faster than traditional defenses. Quilici, a longtime technology entrepreneur, previously co-founded Quack, a startup later acquired by AOL.
From teaching electrical engineering to fighting scam calls, Quilici draws on a career spanning research and real-world security. He notes that AI has made robocalls cheaper and more adaptive, meaning scams now sound natural and are harder to detect in real time.
Scammers quickly adapt by rotating numbers, exploiting weaker carriers, and routing traffic internationally, limiting the impact of call authentication frameworks. Without faster enforcement and real penalties, scammers can keep operating cheaply at massive scale.
Quilici adds that regulators may not have anticipated how difficult it would be to change carrier behavior.
Vishwa: Are there warning signs to distinguish robocalls from legitimate calls? What typically happens during a robocall interaction?
Alex: Yes. Robocalls often create a sense of urgency right away, use generic greetings instead of your name, and ask you to press a key or provide information quickly. Many begin with a pause before the message starts, which is a technical artifact of automated dialing systems.
During a typical robocall interaction, the goal is to keep the recipient engaged just long enough to trigger a response, confirm the number is active, or transfer the call to a live scammer.
Vishwa: If consumers could change their response to robocalls, what should that be?
Alex: The most important change is to not engage at all. Do not answer unknown numbers, do not press keys, and do not respond verbally. Any interaction signals to scammers that the number is valuable.
Let calls go to voicemail and use call blocking tools that rely on network-level data rather than individual reporting alone.
Vishwa: Why do robocalls succeed despite awareness? Which triggers are utilized?
Alex: Awareness helps, but robocalls succeed because they exploit human psychology. Common triggers include fear of financial loss, authority impersonation, such as banks or government agencies, and time pressure. Scammers also rely on volume. Even if a tiny percentage of people respond, the scale makes campaigns profitable.
Vishwa: Can you describe a recent robocall campaign you identified early? What indicators triggered its detection?
Alex: We identify almost all campaigns very early in their lifecycle. For example, there is a Wal-Mart campaign that has been going on for the past year, essentially telling people that there has been a purchase on their account and that they need to press 1 to cancel it.
This campaign was identified early, within the first few calls, because it fit a standard pattern we’ve seen with other retailers, and most of what had changed was the product that was claimed to have been purchased, the cost of the product, and the brand it mentioned.
Vishwa: Can you share steps of how AI changed robocalls? Which trends worry you more than others?
Alex: AI has made robocalls cheaper, faster, and more adaptive. Voice synthesis allows scammers to sound more natural and less scripted. AI also enables rapid testing of different messages to see which ones perform best.
AI allows random messages to be generated, which makes it more difficult to spot the problematic ones. But the most concerning trend is the use of AI to personalize scams at scale, leveraging data from data brokers and data breaches, making calls feel relevant and credible.
Vishwa: What long-term patterns in your robocall dataset concern you most? Which trends are critical?
Alex: One concerning pattern is that, despite new regulations and more enforcement, and despite scam volumes shrinking as regulations increase, the actual dollar losses have increased dramatically. So while they call less, they make more money.
The problem is that scammers adapt faster than enforcement mechanisms. Another critical trend is the globalization of fraud operations, with campaigns launched across borders to evade local enforcement and carrier controls.
Vishwa: Scam and telemarketing calls grew sharply in 2025. What does your data say about these campaigns?
Alex: Our data shows a clear and troubling shift in 2025. Overall robocall volume stayed relatively flat at 52.5 billion calls, declining less than 1 percent from 2024, but unwanted calls together surged more than 15 percent year over year.
These calls now account for roughly 57 percent of all robocalls, up from 49 percent the year before. While legitimate categories like notifications and payment reminders declined sharply, scammers filled the gap with more aggressive and better-targeted campaigns.
They’ve also moved to campaigns that feel like telemarketing but are often illegal in terms of deceiving or defrauding consumers. This tells us the problem is no longer about total call volume alone, but about the increasing concentration of harmful calls that are designed to deceive and steal.
Vishwa: As robocalls become more personalized and AI-driven, what technical tools or resources are leveraged to scam targets?
Alex: Scammers increasingly rely on a combination of leaked personal data, AI-generated voice technology, and automated calling infrastructure. Data from breaches and brokers allows them to reference real names, recent purchases, or financial institutions, which immediately builds credibility.
AI voice systems make calls sound more natural and less scripted, while automated dialing platforms test and refine messages in real time based on which prompts keep people engaged. We still see material use of caller ID spoofing and rapid number rotation, which allows campaigns to scale quickly while staying ahead of traditional blocking methods.
Vishwa: Where are telecom and regulatory defenses failing most? What measurable gaps allow scam volumes to remain stable globally?
Alex: The most obvious gap is that existing defenses are not reducing scam calls at scale. Despite years of regulatory and industry efforts, U.S. consumers still receive more than 50 billion robocalls annually, and nearly 30 billion of those are unwanted telemarketing and scam calls.
Call authentication frameworks help at the margins, but scammers adapt quickly by rotating numbers, exploiting weaker carriers, and routing traffic internationally.
Enforcement remains inconsistent, penalties are slow, and bad actors face little real risk. As long as scam operations can operate cheaply and at massive scale, volume will remain stubbornly high.
Vishwa: Where does the telecom ecosystem still fail to address risks? Which assumptions regulators and carriers still rely on that you suggest might be ineffective?
Alex: The general problem is that carriers make money by carrying call traffic. They also can essentially say they can’t be blamed for the actions of callers, as they are just a pipe and aren’t the police.
While some carriers make efforts to know their customers, monitor the traffic, and eject the problem children, other carriers are happy to essentially look the other way and take in the revenue from those calls.
I think regulators didn’t realize how hard it would be to change the behavior of most carriers, or just how much revenue is likely tied to problematic traffic.
Vishwa: How do robocalls aid financial theft or identity fraud? How does YouMail help?
Alex: Robocalls are often the front door to financial theft and identity fraud. Scam calls impersonate banks, payment services, delivery companies, and government agencies to extract credentials, payment details, or personal information that can later be monetized.
In 2025 alone, nearly 30 billion unwanted telemarketing and scam calls created constant exposure for consumers. YouMail helps by stopping these calls before they ever ring. Our platform analyzes billions of calls using audio fingerprinting, call behavior patterns, and real-time network intelligence to identify malicious campaigns at their earliest stages.
By blocking scams proactively rather than relying on user reaction, we reduce both financial risk and emotional harm for consumers. Not only that, a decent number of carriers are subscribing to YouMail services we provide that inform them when our consumers were hit with illegal traffic they initiated, allowing them to shut off those customers.
Related
Get expert insights on threats, breaches, scams, and security trends — delivered every Monday.
Most Popular
Get expert insights on threats, breaches, scams, and security trends — delivered every Monday.
Most Popular