Battle plans to tackle Ad Fraud: Can AI Stay One Step Ahead? Part - 1

The loss is staggering! Twenty-two per cent of all online ad spend – totalling $84 billion – is swallowed by ad fraud, per a 2023 study by Juniper Research. In just five years, that number is expected to double, exceeding $170 billion. Eliminating ad fraud entirely is as unlikely as stamping out shoplifting for good, but in this intense battle, Artificial Intelligence has emerged as a powerful ally.

Yet, AI is a double-edged sword: it is both the most promising tool to fight ad fraud and, paradoxically, a resource that fraudsters themselves use to sidestep detection. As advertisers turn to AI to analyse patterns, detect anomalies, and flag fraud proactively, fraudsters are also employing AI to launch more sophisticated, elusive attacks. This dual nature raises a critical question: can AI’s fraud-detection capabilities keep pace with the innovation of those who misuse it?

In this two-part series, Adgully examines the fine line between AI’s potential to strengthen ad defences and the relentless challenge of outsmarting those determined to exploit its vulnerabilities.

Honestly, it is an endless chase, quips Kruthika Ravindran, Director, Key Accounts, TheSmallBigIdea. On the one hand, she adds, its enhanced techniques can track fraudsters who would otherwise pass undetected by human vigilance. “But those very fraudsters are also mastering the tools and technologies of AI to evade the very barriers that have been put in their way! Luckily, AI can learn and adapt. There can be a proactive defence mechanism. By continuously learning from new data, AI can become increasingly adept over time, improving predictive accuracy and adapting to evolving fraudulent tactics,” she says.

Dr Ankoor Dasguupta, Chief Experience Officer, Garage Collective, sees AI as a double-edged sword in the context of ad fraud. According to him, before aiming for balance, the first priority should be reducing the widespread occurrence of ad fraud.

“Regular audits are essential, along with a focus on continuously refining AI models and algorithms. Organisations should employ adaptive learning techniques that not only detect known fraud patterns, but also anticipate emerging tactics. To stay ahead of increasingly sophisticated fraudsters, a multi-layered approach is advisable – combining AI-driven solutions with human oversight to ensure any anomalies flagged by AI are thoroughly investigated and verified,” he states.

Dasguupta notes that beyond natural language processing (NLP), several notable advancements in AI-powered tools are playing a pivotal role in the fight against ad fraud. Anomaly Detection Algorithms, for instance, leverage machine learning to identify deviations from typical user behaviour, which can signal fraud. “Behavioural Biometrics is another advancement, analysing traits like typing patterns and mouse movements to better determine whether interactions are authentic or fraudulent. Building trust requires what I see as a non-negotiable, on-going dialogue between advertisers and AI developers. Regular engagement through forums, workshops, and collaborative projects ensures that solutions remain effective and aligned with the evolving challenges of ad fraud.”

Aakash Goplani, Account Director, SoCheers, suggests that advertisers should make the robust use of AI in detecting ad frauds.

“AI plays a dual role in the advertising world. On the one hand, its advanced algorithms are instrumental in analysing vast amounts of data and identifying fraudulent activity, thus improving campaign efficiency. On the other hand, fraudsters are leveraging AI to enhance their deceptive tactics, creating a challenging dynamic. Advertisers and tech companies must invest in robust AI solutions that not only detect existing fraud patterns, but also evolve to counter new deceptive methods. This proactive defence is essential in this changing landscape, to stay ahead in this cat-and-mouse chase,” he adds.

AI is a powerful technology that can be wielded for both constructive and destructive purposes, much like a hammer, says Vishal Rupani, Co-founder, Sprect.com.

“With a hammer, one can create beautiful furniture or art, but it can just as easily be used to smash a glass. The responsibility lies with the user – if they choose to use AI for good, such as developing tools that enhance productivity or improve user experiences, the results can be transformative. Conversely, if misused, AI can facilitate malicious activities like ad fraud or misinformation campaigns. For example, someone could use AI to generate a script that automatically clicks on the first ad in a Google search for a specified keyword. Illustration below shows how AI can be harnessed for unethical practices, depending on the intent behind its application,” says Rupani.

This example, according to Rupani, highlights how powerful AI tools can be misappropriated, emphasizing that while technology itself is neutral, its impact is determined by how we choose to use it.

“Imagine a game of cat and mouse, but with super-smart mice that have jetpacks! The catch? The cat (our AI) has to keep getting smarter to catch the mice (the scammers). It’s like a never-ending chase, with the bad guys getting sneakier and the good guys trying to stay one step ahead. So, while AI can help fight ad fraud, it also gives the bad guys new toys to play with,” he adds.

Shan Jain, an independent director, brand strategist, and marketing transformation advisor, sees AI’s role in ad fraud as similar to an evolving immune system. While AI fortifies our defences, fraudsters are constantly adapting to exploit new vulnerabilities.

“Just as an immune system fights viruses, AI identifies and blocks ad fraud by detecting irregular behaviour and swiftly neutralizing threats, with real-time pattern recognition as its core strength. However, fraudsters, like viruses, are evolving too. They now use AI to create increasingly sophisticated fraud schemes, making this a continuous battle where AI must constantly refine its immune response,” Jain adds.

Russhabh R Thakkar, Founder and CEO, Frodoh World, states that AI plays a dual role in the ad fraud landscape. As a solution, he points out, it significantly enhances our ability to detect and prevent fraudulent activities by processing vast amounts of data in real-time and identifying complex patterns.

“However, it also presents a challenge as sophisticated fraudsters leverage AI to create more convincing fake traffic and evasion techniques.The key to maintaining a positive balance lies in continuous innovation and the integration of AI with human expertise. While AI provides the speed and scale needed to combat fraud in the digital ecosystem, human insight remains crucial for contextual understanding and strategic decision-making,” he adds.

(Stay tuned for Part 2 tomorrow, where we’ll dive into the latest advancements in AI-powered tools that are proving most effective in combating ad fraud. We’ll explore how these tools stay ahead of increasingly sophisticated fraud tactics and examine the critical role of trust in AI's role in ad fraud prevention.)