Are We Truly Safe As Generative AI Fuels a Surge in Crime?

Are we truly safe when Generative AI is continuously increasing criminal activities? This technology has risen as a life-changing force in AI and can generate unlimited texts, images, music, and much more. Also, human input is very low during the production of the above-mentioned things because now technology has the power to change the field ranging from content generation to automated businesses. ChatGPT is one of the best examples of generative models—it can produce many blogs, create designs, and conduct human-like conversations. 

However, several tools can produce the artwork depending on the prompt text.  All these proficiencies indicate the huge potential of generative models to increase productivity and change across different industries. Furthermore, with technological advancement, crimes are also increasing because generative AI can produce highly realistic yet fake visuals of someone that provide the chance to others of its misuse. Does it mean our security is at risk because of the wrong use of AI which is influencing bad on society? 

Rise of Generative AI in Crimes

Generative AI has changed fraud ways and enabled criminals to bring their alternative methods to the market against financial organizations and consumers. The only limit is how criminals are thinking to conduct a crime while using generative AI deepfakes for clever scams. However, these deepfakes use self-instructed systems that constantly adopt the old detection techniques. Producing fake videos, voices, and even documents has become easy by using fake yet cheap actors while using generative models. Dark webs are one of the common examples that are filled with fake content and fraudsters can get access to these sites for just $20 for huge attacks. 

The increase in generative AI fraud is striking, with deepfake incidents in the fintech sector skyrocketing by 700% in 2023, prompting financial institutions to urgently seek effective defenses. Business email compromises are especially at risk, resulting in nearly $2.7 billion in losses in 2022. Generative models enable fraudsters to target multiple victims at once, which could lead to losses surpassing $11.5 billion by 2027. Although banks have traditionally been at the forefront of combating fraud, their current risk management strategies often struggle to keep pace with the new AI-driven threats. 

Major Types of AI-Driven Violation

Let’s discuss some of the incidents that highlight the quick need for robust rules and proper levels for the use of generative AI to defend institutions’ privacy and reduce the misuse of generative models across different industries: 

Facebook and Cambridge Analytica (2016): A quiz on Facebook collected personal information that Cambridge Analytica used to develop targeted political advertisements. This data misuse sparked privacy issues and led to Facebook facing a $5 billion fine.

Strava’s Heatmap Issue (2018): The fitness app Strava exposed users’ workout locations, which included private addresses and sensitive areas such as military bases. This occurred because the default user settings permitted data sharing.

Dinerstein v. Google (2019): Google is being sued for allegedly accessing patient data without consent to train its AI. This case highlights the ease with which personal information can be misused.

IBM’s Facial Recognition Problems (2020): IBM utilized images from the internet without obtaining permission to compile a dataset aimed at enhancing AI recognition of various skin tones. This resulted in a lawsuit for breaching privacy laws.

Clearview AI’s Database (2020): Clearview AI gathered millions of images from social media to create a facial recognition database, which it then sold to law enforcement agencies. This led to legal challenges, emphasizing the importance of obtaining consumer consent.

Aon’s Biased Hiring Software (2023): Aon’s hiring assessments have been criticized for discriminating against specific groups. The ACLU has complained, highlighting the importance of fair practices in companies that utilize generative AI. 

Real-Life Case Due to AI Scams

The article in the Wall Street Journal exposed the fraudster’s mechanism for manipulating the generative AI for deepfake fraud. Scammers generate realistic yet fake visuals to mock people as victims and cause their financial losses. Guo is one of the most shocking cases who mistakenly transferred up to $600,000 after a fake video call that was acting like his friend. The truth is revealed when Guo meets his real friend. 

So, these events highlight the quick and robust requirement of deepfake fraud detection techniques to fight against such scams. Moreover, the generative models’ emergence is making old fraud prevention methods incapable and encouraging governments to seek regulations worldwide. The enhanced authentication and immediate monitoring will act as strong security techniques and be helpful against the latest fraud methods. Besides, financial institutions should connect with the high authorities to generate a proper structure that promotes innovations that reinforce the security against AI-driven scams.  

Impacts of AI-Generated Frauds on Society