Google says it has fixed the image generator that wasn’t able to properly depict white people.

Introduction

Over the past few years, there has been growing alarm over the depiction and presentation of multiple races on various media platforms. The recent wave of artificial intelligence (AI) technology has once again raised questions surrounding images and how we depict white people. Most recently, Google — a tech giant if there ever was one—saw widespread outrage when an image generator AI they built ended up not representing white people correctly. This is something Google has worked to address in its AI system — trying not only to make it fairer, but inclusive too.

For more articles check  Newsworldportal

Background

Today, however, artificial intelligence powers every technological interaction and output we engage with. Google has been leading in creating AI powered solutions that deliver users new and interesting experiences. AI: Computer vision, which is an important area of AI where machines are designed to understand and interpret visual data from the real world like pictures(images) or videos.

Google’s Image Generator

An example of a computer vision technology development is an image generator that Google created capable to generate very realistic pictures with only descriptions, or keywords. For instance, a developer can provide criteria — age, gender or race of the people in an image and AI will build that specific picture.

The Problem with White Characters

With that said, some users found out recently they could confuse Google’s image generator because it was apparently struggling to render white people correctly. The images created by such systems would often have unnatural or twisted features, which raised concerns about racial bias and inclusion. These deficiencies spurred a discussion of the inherent biases in AI systems and how they can propagate through society.

Google Solution for The Image Generator

Google responded quickly, acknowledging the gravity of their image generator AI failure. They started to immediately work on strengthening the system and how they would provide an honest approach with all races — including whites.

A team of professionals from Google introduced extensive testing and fine-tuned numerous algorithms to ensure representation was equal among various demographics. They also included training data for a larger number of subjects, which ultimately helped the AI to learn and adapt to a broader set of human characteristics. This was a conscious effort to correct for some biases of the original image generator, and presented as an inclusive AI environment.

Discussion on the Impact

Depicting the consequences of proper representation in AI ecosystems As AI continues to be used in places like marketing, entertainment or even policing as well it becomes highly pertinent to ensure that the systems do not end up embedding racial discrimination through its uses. This is in turn Google addressing problems with its image generator and contributing to a better future of equal access for every person — the core values on which AI technology revolves.

Nonetheless, the incident is an illustration of how biases can become embedded into AI systems in general. The algorithms and machine learning techniques that are necessary to power these smart systems may, in turn, perpetuate implicit biases. Developers and researchers, on the other hand, must always be alert to work for transparency as fairness or take into consideration that AI systems need allow diversity in their implementation.

Conclusion

A strong signal from Google to fix race and gender bias in Alphabet’s Machine Learning system is clear when the image generator gaps being learned. Google deserves credit for admitting the issue and making an effort to correct it, but they are far behind where other tech giants should be in realizing this. As we see AI technology grow, it is incumbent upon companies to take concrete steps in making the algorithms that they use diverse enough so as not to lead us into a future where persons of color have been bred out by algorithmic exclusion. This is the only way that we can really unlock AI to help build a more equitable and integrated society.