Michelle ThompsonBooking a trip online involves placing significant trust in a platform—trusting it with your money, personal information, and travel plans. For a company as large as Booking.com, maintaining that trust for millions of users daily is a major responsibility. Increasingly, artificial intelligence is playing a central role in keeping that trust intact.
Combating today’s online fraud demands the use of AI.
The volume of data Booking.com manages is staggering. Protecting users goes far beyond preventing the use of stolen credit cards; the platform also has to detect fake hotel reviews, marketing scams, phishing attempts, and account takeovers. “We apply AI across a wide range of safety and fraud prevention tasks,” said Choudhury. “
This involves handling petabytes of data, including events from applications, infrastructure, messages, emails…”
To manage this complexity, they don’t rely on a single solution. “We combine multiple vendor-specific machine learning tools with our own in-house systems to identify and stop fraudulent activity,” Choudhury explained.
In essence, Booking.com merges off-the-shelf software with custom-built AI, creating a robust security system that safeguards both travelers and property owners on the platform.
The million-dollar question: better or cheaper?
Naturally, managing a security system at this scale is no small feat. One major challenge is simply ensuring that all the different tools—both in-house and third-party—work together seamlessly. Yet, Choudhury highlights an even more difficult, ongoing trade-off: balancing performance with cost.
Cyberattacks are evolving constantly, requiring defenses to improve just as quickly. However, enhanced technology comes at a higher price.
“With cyber threats becoming more sophisticated and the volume of data continuously growing,” Choudhury explains, “we face a choice: do we prioritize cost-efficiency, or do we focus on achieving the highest performance?”
How AI Stays One Step Ahead of Online Fraud?
As the saying goes, the best defense is a strong offense. Rather than reacting to issues after they occur, Booking.com leverages AI to identify potential problems before they arise. A key part of this strategy involved migrating their systems to the cloud, enabling faster and smarter tools.
Choudhury noted that human security experts now have digital teammates. “Multiple AI assistants work alongside security analysts to boost efficiency and reduce operational burden,” he said.
By pairing top human investigators with AI that can sift through massive amounts of data in seconds, experts can concentrate on the most critical threats, while robust monitoring ensures the AI continues to operate accurately and reliably.
I can continue paraphrasing the next sections in the same professional style if you want.
Ensuring AI Operates Responsibly
When AI is entrusted with critical security decisions, it’s essential to ensure it operates fairly. Choudhury emphasized that ethics are central to Booking.com’s approach, which rests on several key principles:
- Fairness: The company regularly audits its AI to prevent biases, ensuring no individual or group is unfairly targeted for potential fraud.
- Human Oversight: Major decisions always involve human review. “There is enough human involvement to identify false positives,” Choudhury explains—final decisions are often made by people, not just algorithms.
- Explainability: AI cannot remain a mysterious ‘black box.’ The team needs clear insight into why the system made specific decisions to maintain accountability.
- Privacy: All systems are grounded in robust data protection, privacy, and compliance measures, with user consent always a priority.
What the Future Holds?
Choudhury sees the next major advancement not in creating entirely new AI applications, but in making existing AI tools work together more effectively.
“I anticipate more solutions being developed, but the real impact will come from orchestrating them to make departments significantly more efficient,” he predicts.
The aim is to create a system where all security components communicate and collaborate intelligently. For Booking.com, the focus is clear: “Driving innovation while ensuring reliability and cost efficiency will remain the top priority.”
For everyday users, this means AI is helping provide extra reassurance that online bookings are secure from fraud.
Siddhartha Choudhury and the Booking.com team will share further insights at this year’s AI & Big Data Expo Europe in Amsterdam on 24–25 September 2025. Choudhury will participate in a panel titled “Innovation at Scale: Gen AI, Cloud Platforms, and Data-Driven Development” on the second day of the event.
Emily Martinez
