LinkedIn, a leading tech giant, has recently collaborated with the University of California, Berkeley to conduct groundbreaking research on AI image detection. The objective of this research is to achieve an impressive 99% success rate in detecting fake profile images, while maintaining a mere 1% false positive rate. The evidence presented by the researchers indicates that this endeavor has yielded promising results.
According to a top engineer at LinkedIn, the online information reveals that the app's new photodetection tools have significantly reduced the number of false positives, resulting in higher rates of accurate positive identifications. This development has sparked discussions among individuals regarding the reasons behind creating such profiles on the platform.
One possible motive is the belief that it could benefit the community through improved search engine optimization (SEO), making Google appear more trustworthy by associating author biographies with profiles on the app. Additionally, some individuals may be motivated to create websites that are perceived as more credible to benefit others. Lastly, there exists a small segment of individuals driven by the desire to establish a highly reliable page where visitors can feel secure. However, it is crucial to highlight that engaging in such practices is ethically unacceptable.
The realm of AI technology is currently striving to generate the best profile pictures. However, this pursuit can inadvertently lead to the proliferation of fake profiles, exacerbating an already significant issue. Reports published recently shed light on how LinkedIn successfully detected and eliminated over 21 million fake accounts during the early months of 2022.
LinkedIn employs a method to identify AI-generated content by identifying structural differences present in such images. Unlike genuine images, artificially created images exhibit consistent patterns that set them apart.
To illustrate this distinction, LinkedIn presented an example consisting of a composite image comprising 400 artificial images and another composite comprising 400 real images. The composite of the fake images demonstrated noticeable similarities in the regions surrounding the eyes and nose. Conversely, the composite of the real images exhibited no commonalities with any of the other images, resulting in a blurred composite.
Affiliate marketers have also provided substantial evidence, recounting their experiences of creating fake profiles on the app. These practices can now be circumvented through the utilization of LinkedIn's AI image detection tools, which is expected to have a considerable impact on reducing the prevalence of fake accounts.
LinkedIn acknowledges the importance of continuously improving and enhancing the effectiveness of their measures to combat abuse and malicious behavior. As part of their efforts, the company has forged partnerships with academic institutions, positioning themselves at the forefront of innovation in this field.
SO, WHERE DO YOU FIND THIS PARTNER?
Well, aren’t we glad you asked! We at DigiCom are obsessive data-driven marketers pulling from multi-disciplinary strategies to unlock scale. We buy media across all platforms and placements and provide creative solutions alongside content creation, and conversion rate optimizations. We pride ourselves on your successes and will stop at nothing to help you grow.
Comentarios