[By André Ferraz, CEO and Co-Founder of Incognia, the innovator in location identity solutions]
Generative artificial intelligence (GenAI) is a hot topic of conversation – particularly the risks it poses to users’ online safety and privacy. With President Biden calling on Congress to pass bipartisan data privacy legislation to accelerate the development and use of privacy-centric techniques for the data that is training AI, it’s important to remember that excessive regulation can stifle experimentation and impede the development of new and creative solutions that could change the world.
For example, in the early days of telecommunications, regulatory bodies had a strong grip on the industry, yielding monopolies via limited competition. It wasn’t until the U.S. government took anti-trust action in the 1980s that deregulation led to waves of innovation, including the development of technologies that accelerated the adoption of the internet, mobile phones, and broadband services.
This isn’t to say that GenAI shouldn’t be regulated, it’s just important to differentiate between hypothetical doomsday scenarios associated with AI and the real-world impact of technology today. The concern with large language models (LLMs) and GenAI is the potential for misuse by bad actors to generate harmful content, spread misinformation, and automate and scale malicious or fraudulent activities. Any regulations or new technologies introduced should be focused on addressing these particular challenges.
Worried About the Wrong Things
There is often a focus on events that may or may not materialize in the distant future – with governments and citizens worldwide expressing concerns about the potential for AI to become uncontrollable, leading to unforeseen consequences. This fixation on existential risks diverts attention from the immediate and tangible challenges AI poses today – namely, the profound implications of GenAI on fraud prevention strategies and user privacy. While it’s essential to anticipate and address a variety of long-term possibilities, it’s equally vital to concentrate on the real-world impact of GenAI in daily life at the present moment.
AI-Driven Benefits
AI has already demonstrated its capacity to revolutionize industries and improve various aspects of our lives. From healthcare, where AI aids in early disease detection, to autonomous vehicles that enhance road safety, AI’s contributions are promising. Additionally, AI has already become an integral part of our digital lives through voice assistants that streamline tasks and improve accessibility to recommendation algorithms that personalize our online content.
Overregulation across use-case contexts can hinder the development of new and creative solutions that could benefit society. We need to strike a balance, recognizing that while risks are indeed present, AI’s potential for good is immense, and our focus should be on harnessing this potential responsibly.
Stopping the Bad Actors
One of the most immediate and pressing concerns in the GenAI landscape is the misuse of the technology by malicious actors, and the looming threat posed by rapidly advancing artificial intelligence cannot be ignored. Today, numerous industries, including financial institutions and online marketplaces, heavily rely on document scanning and facial recognition technologies for robust identity verification protocols. However, the stark reality is that the proliferation of deepfakes coupled with GenAI capabilities has rendered these traditional methods increasingly vulnerable to exploitation by fraudsters.
Facial recognition, once considered a reliable authentication tool, is now susceptible to manipulation. Our digital footprints, readily available on social media platforms and various online databases, serve as fodder for fraudsters to craft sophisticated masks, circumventing facial recognition systems with ease. Even liveness detection mechanisms, previously hailed as a safeguard against impersonation, have been compromised by the advancements in GenAI.
The reliance on publicly available information for identity verification is proving inadequate in thwarting fraudulent activities. While these methods may superficially fulfill compliance requirements, they fall short in effectively combating fraud.
The proliferation of data breaches has rendered personally identifiable information (PII) essentially public domain, further undermining the efficacy of conventional identity verification techniques. Document verification, facial recognition, and PII authentication are all vulnerable in the face of GenAI’s evolving capabilities.
Addressing this challenge requires a multifaceted approach that goes beyond regulatory measures, and focuses more on the introduction of new technologies. Pattern recognition, for instance, plays a pivotal role in identifying abnormal and potentially harmful behaviors. By training GenAI models to recognize patterns of behavior associated with malicious intent, it can swiftly identify and respond to potential threats.
Real-time detection is another essential component in spotting, stopping, and combating bad actors or harmful generated content and activities. AI systems can monitor user behaviors during interactions and transactions, flagging suspicious activities and allowing for immediate intervention, thus preventing or mitigating potential harm. Additionally, user behavior profiling can provide valuable insights into identifying malicious actors, anomalies, and potential threats. By creating detailed profiles of typical user behaviors, AI systems establish a baseline for normal behavior and quickly flag deviations that may indicate fraudulent intent or harmful actions.
GenAI is a Dynamic Field
GenAI holds vast potential to reshape industries, drive innovations, and improve various aspects of our lives. However, with great power comes great responsibility, and AI is no exception. While responsible and effective regulation is essential, it’s important to avoid overregulation that could impede progress and innovation.
The challenge posed by GenAI is not merely transitory; it represents the future landscape of fraud prevention. Organizations relying solely on these traditional approaches for Know Your Customer (KYC) and Identity Verification (IDV) procedures must urgently reassess their strategies.
Additionally, businesses in the fraud prevention space shouldn’t expect users to protect themselves. To tackle the real dangers of AI, a targeted approach is needed – leveraging solutions that prevent GenAI abuse, to protect users and their data. This nuanced strategy considers the diverse risks and benefits of various AI applications and, instead of adopting a one-size-fits-all approach to regulation, can consider the multifaceted nature of the emerging technology. The future of AI regulation should strike a balance between safeguarding ethical practices and fostering creativity and progress in the AI landscape.
Companies should invest in fraud prevention solutions that use GenAI to find data points that allow them to more uniquely identify their users with a proactive approach, being the first risk signal to determine the misuse of GenAI.
