Uncovering the Secrets of Google AI: A Photo Analysis Breakthrough

Background on Ente

Ente is a photo-sharing startup founded by Vishnu Mohandas, a former Google engineer. He created the service as an alternative to Google Photos, which he feared could be used to train AI systems that might be used for purposes he didn’t control.

How Ente Works

Ente uses computer vision to analyze uploaded images and provide detailed descriptions of them. The website also allows users to experiment with analyzing pre-chosen stock images.

Comparison with Google Photos

Google’s computer vision is able to pick up on subtle details in the stock images, such as a person’s tattoo or a child’s temporary tattoo. However, Mohandas notes that if you don’t want to upload your own picture, Ente gives people the option to experiment on Theyseeyourphotos using several pre-chosen stock images.

Concerns about Google Photos

Mohandas is concerned that humanity’s visual archive will be mined in the future in ways he can’t predict or control. He fears that photos snapped of his daughter today could be used to manipulate her decades from now by anyone who has access to this data, including advertisers, dating websites, employers, and industries that don’t exist yet but will benefit from psychological profiles.

Google’s Response

A Google spokesperson declined to comment directly on Ente’s project. However, the company states that uploads to Google Photos are only used to train generative AI models that help people manage their image libraries, like those that analyze the age and location of photo subjects. The company says it doesn’t sell the content stored in Google Photos to third parties or use it for advertising purposes.

Transitioning from Google Photos

If you want to switch from Google Photos to another image storage service, the transition might not be totally smooth. Mohandas notes that Google makes it difficult for people to transfer their photo library elsewhere by breaking up files and compressing them.

Key Points

• The website uses computer vision to analyze uploaded images and provide detailed descriptions of them.

• Google’s computer vision can pick up on subtle details in stock images, but Mohandas notes that the service gives people the option to experiment with analyzing pre-chosen stock images.

• Mohandas is concerned about the potential for humanity’s visual archive to be mined in the future in ways he can’t predict or control.

Google’s AI and Photo Sharing: A Concerning Relationship

Introduction

The Concerns

Vishnu Mohandas, an ex-Google engineer, founded an alternative service called Ente, which is open source and end-to-end encrypted. He was concerned that his content could be used to train AI systems, even if they weren’t specifically tied to the Pentagon project. This concern led him to develop a website called Theyseeyourphotos.com, where users can upload photos and see what Google’s AI models can learn from them.

How It Works

The website uses a Google Cloud computer vision program that writes a three-paragraph description of the uploaded images. The program documents small details in the images, including objects, people, and their actions. However, this raises concerns about how much information is being collected and used by Google’s AI models.

A Concerning Example

Mohandas tried uploading a photo of his family to Theyseeyourphotos.com. The AI model noted that his wife was wearing a Casio F-91W watch, which is commonly associated with Islamic extremists. This example highlights the potential for bias in Google’s AI models and the need for transparency.

A Warning from Mohandas

The Alternative Solution

Ente offers a paid service that is open source and end-to-end encrypted. It has over 100,000 users, many of whom are already part of the privacy-obsessed crowd. However, Mohandas struggled to articulate why people should reconsider relying on Google Photos despite its conveniences.

Conclusion

The relationship between Google’s AI and photo sharing raises concerns about data collection and use. While Google claims that it doesn’t sell user content or use it for advertising purposes, the extent of their data collection is unclear. The website Theyseeyourphotos.com provides a glimpse into what Google’s AI models can learn from user photos, highlighting the need for transparency and regulation in the industry.