AI, Internet Scams, and the Balance of Freedom | Chris Olson | EP 460

Added: Jul 2, 2024

Vulnerable Groups and Common Scams

Olson explains that elderly people, particularly those in their late 70s to mid-80s, are frequently targeted by romance scams initiated on dating websites. Criminals collect data on seniors, interact with them, and then move conversations off the dating sites to manipulate them emotionally and financially. The scammers often have sophisticated profiling methods to identify vulnerable targets.

Another common scam targeting seniors involves fake tech support or computer upgrade offers. Seniors often receive pop-ups or messages claiming there's an issue with their device, prompting them to call a number or click a button, leading to further exploitation.

Olson mentions that 17-year-old males are often targeted with offers for illicit drugs, including a growing market for under-the-table hormonal treatments. Fourteen-year-old girls interested in modeling are frequently approached by human traffickers. The sick and infirm are targeted with false medical offers and scam products, especially prevalent during the COVID-19 pandemic.

Scope of the Problem

Olson estimates that for a typical elderly woman between 78 and 85 years old, about 2.5% to 3% of every page impression or app view attempts to target them with some form of crime or harmful influence. He notes that if grandmothers were being mugged or attacked on the street at the rate they're targeted online, there would likely be National Guard presence on every street in America.

Third-Party Code and Content Delivery

Olson explains that roughly 80% of what runs on a user's device when visiting a website or using an app comes from third-party sources, not the owner of the site or app. This third-party code is where much of the criminal activity originates. The Media Trust Company's job is to examine this third-party content, determine what violates company policies or might harm consumers, and inform companies about these violations.

AI and Deep Fakes

The conversation touches on the growing use of AI in creating deep fakes and manipulating content. Olson mentions that while the fear of deep fakes is currently greater than their actual use in paid content, the technology is rapidly evolving. He notes that criminals are now creating multiple copies of similar content to make it harder to remove entirely.

Data Collection and Targeting

Criminals leverage the vast amount of data collected by digital companies to target individuals effectively. This data allows them to create highly personalized scams and attacks. Olson emphasizes that the digital ecosystem knows a great deal about users, including seniors, making it easier for criminals to exploit vulnerabilities.

Challenges in Combating Digital Crime

Peterson points out the difficulty in prosecuting online criminals who often operate from different countries. Olson acknowledges this challenge and suggests that cooperation between governments and private entities in the digital supply chain is crucial for addressing the issue effectively.

While education about online dangers is important, Olson argues that it's not sufficient, especially for seniors who may lack the technical savvy to distinguish between legitimate and fraudulent communications. He emphasizes the need for technical solutions and cooperation between governments and private companies.

Olson discusses the challenge of balancing user privacy with the need to monitor and prevent criminal activity. He suggests that while privacy is critical, there needs to be a way to understand and control what is delivered to users' devices based on collected data, without necessarily broadcasting that information.

Potential Solutions and Approaches

Olson advocates for governments to take a more active role in policing digital crime, similar to how they police physical streets. He suggests that state governments, being closer to the people, might be better positioned to take tactical action against digital crime.

He emphasizes the need for cooperation between governments and private companies in the digital ecosystem. He suggests that companies should be willing to share information about criminal activities without fear of repercussions, to collectively move upstream and address the source of harmful content.

Olson argues that there needs to be a shift in focus from protecting corporate assets to protecting consumers. He suggests that companies prioritizing digital safety and consumer trust are likely to be more successful in the long run.

He explains how his company creates millions of virtual personas that mimic potential targets to attract and identify criminal activity. These personas are designed to appear as vulnerable individuals to the digital ecosystem, allowing the company to detect and report harmful content and activities.

While AI is increasingly used by criminals, Olson notes that it's also being employed for protection. His company uses AI to continuously scan digital assets and keep their virtual profiles active, essentially using AI to defend against AI-enabled criminal activities.

Economic Implications

Peterson suggests that companies prioritizing user safety and trust are likely to be more successful in the long term. Olson agrees, noting that companies willing to engage with governments to combat bad actors will play a crucial role in improving the digital ecosystem.

He points out that some companies are willing to withhold significant revenue to prevent the possibility of targeting something harmful. However, he notes that many companies are not willing to go that far, especially in the current "risk-on" phase in the digital world where the race for technological supremacy often overshadows safety concerns.

Future Outlook

Olson expresses cautious optimism about the ability to control online criminality. For companies that prioritize digital safety, he believes it's possible to detect and remove the vast majority of potential attacks. However, he acknowledges that it's not possible to eliminate the problem entirely due to the nature of the internet and data targeting.

The conversation concludes with a discussion of the ongoing challenges in the digital world. He notes the current focus on winning the AI and language model race, which sometimes overshadows safety concerns. He emphasizes the need for companies to balance technological advancement with safe implementation.

Conclusion

The podcast highlights the pervasive nature of online criminality and the particular vulnerability of certain groups, especially seniors. It underscores the complex challenges in combating digital crime, including jurisdictional issues, the rapid evolution of technology, and the need to balance privacy with protection. The conversation emphasizes the importance of cooperation between governments and private companies, the potential of AI and virtual personas in fighting crime, and the long-term benefits of prioritizing user safety and trust. While the challenges are significant, the discussion suggests that with the right approach and priorities, it's possible to make substantial progress in creating a safer digital ecosystem.