AI Against Itself: Arjun Singri On Why The Future Of Scam Prevention Depends on Intelligent Defense

As digital communication becomes a routine part of everyday life, from banking alerts to delivery updates and personal messages, cybersecurity experts say the risks people face online are also evolving at the same pace. Artificial intelligence is now not only powering innovation but also reshaping how scams are created and delivered.

For Arjun Singri, founder of Scamless, this has created a new kind of digital arms race, where the same technology driving progress is also being used to deceive people at scale.

Singri positions Scamless as part of a growing response to this shift. The platform is designed to detect and interpret potential scam signals across digital conversations, offering users real-time insight into whether a message may be misleading or manipulative.

“Artificial intelligence has changed the structure of scams entirely,” he says. “It allows bad actors to create messages that feel precise, contextual, and emotionally convincing. The volume increases, the quality improves, and the line between genuine and fraudulent becomes harder to detect.”

He points out that many people still rely on outdated assumptions, such as looking for spelling mistakes or awkward phrasing, to identify scams. But today’s AI-generated messages can closely mimic trusted brands, friends, or institutions, making them far more difficult to spot.

That shift is particularly important for everyday users who are constantly receiving messages across email, SMS, and messaging apps throughout the day.

“People assume awareness is enough,” Singri says. “Awareness matters, but it does not scale with the speed of communication. You cannot expect individuals to evaluate every message or link with perfect judgment every time.”

He explains that modern scams are designed to take advantage of how people actually behave online. Most decisions are made quickly, often while multitasking, and under subtle pressure.

“Scams succeed because they are engineered around human behavior,” Singri says. “They introduce urgency, they create a sense of authority, or they trigger emotional responses. When someone is asked to act quickly, the ability to analyze drops. That is where mistakes happen.”

He adds that vulnerability is not limited to any one group. Even experienced or cautious users can be caught off guard if a message appears to come from a trusted source at the right moment.

“This is what makes the problem so widespread,” he notes. “It is not about intelligence or experience. It is about timing and context.”

Within this environment, Singri believes artificial intelligence must also be part of the solution. If scams are now being created and refined using AI, he argues, then detection and prevention must operate at the same level.

“It is no longer realistic to rely on manual recognition alone,” he says.

Scamless uses machine learning systems to analyze patterns in communication that may signal manipulation. These include subtle changes in tone, inconsistencies in language, and behavioral markers that are difficult for humans to notice in real time.

“The goal is to give users clarity in the moment,” Singri says. “Instead of asking them to guess, the system explains what it sees and why it matters. That changes how people respond.”

He describes this as a way of supporting, rather than replacing, human judgment. By adding an extra layer of analysis in real time, users have a chance to pause and reconsider before taking action.

“People do not need to become cybersecurity experts,” Singri explains. “They just need support at the exact moment a decision is being made.”

Singri also highlights a growing imbalance between large organizations and individuals. Companies often rely on advanced security systems to monitor threats and protect communications, while everyday users are left with far fewer protections in similar environments.

“Enterprises have layers of protection built into their systems,” he says. “Individuals are navigating the same risks without that infrastructure. That creates an uneven playing field.”

As AI continues to improve the sophistication of scams, he believes this gap could widen unless more accessible defense tools become available to the public.

For Singri, Scamless represents part of that shift. The aim is to bring more advanced detection capabilities into everyday communication channels where most people actually interact, such as messaging apps, email, and online platforms.

“We are building something people can trust in real time,” Singri says. “It has to be simple to use, but powerful in what it can detect. That combination is what creates confidence.”

He adds that the future of digital trust will depend on systems that can evaluate intent and risk at scale, not just after a scam has already occurred.

“Human awareness still matters,” he says, “but it cannot carry the full weight of modern digital risk. The future of safety lies in intelligent systems that work alongside users, continuously evaluating the environment and providing context when it matters most.”