Imagen 4 Launched: Google's New Weapon in the AI ​​Image Generation Race

In the relentless race between technology giants to dominate the artificial intelligence market, Google continues to demonstrate its ambition by announcing the latest image-generating AI model called Imagen 4. The launch of Imagen 4 and especially the premium version Imagen 4 Ultra is expected to elevate the quality of AI images to a new level, especially in the ability to display text, an inherent weakness of many current image-generating AI tools.

In parallel, Google also made important adjustments to the privacy and operation of the Gemini tool, a move that immediately attracted the attention of the global technology community. The article below will help you better understand two key developments of Google: upgrading the image creation tool and changing user privacy with Gemini.

1. Google Imagen 4 Model

1.1. Image 4

According to Google, Imagen 4 was developed with the goal of becoming the “first choice for most tasks” AI image creation tool. This is a confident statement, showing that Google not only wants to compete but also surpass strong rivals such as Dall-E (from OpenAI) or Midjourney - names that are currently dominating thanks to their ability to create creative, detailed images that are close to human artistic style.

At $0.04 per image, Imagen 4 is considered suitable for general image creation needs, from article illustrations, communication design, to experimental creative projects. Google emphasizes that the text display has been significantly improved compared to the previous version Imagen

3, which was complained about distorted or incorrect text.

1.2. Image 4 Ultra

Along with Imagen 4, Google also launched Imagen 4 Ultra, which is positioned as a premium package for professional users who require the highest level of accuracy and prompt compliance. Google claims that Imagen 4 Ultra is capable of “more accurate text prompt tracking” and produces “stronger” images than competitors such as Dall-E or Midjourney.

Of course, with the quality comes a 50% higher price tag, or $0.06 per image. According to Google, this reflects the high level of customization and detail, making it ideal for professional projects like comics, book covers, animations, product placement, or anything else that requires “on-script” images.

1.3. Good quality, but not as "hot" as before

Objectively speaking, Imagen 4 has made a clear technical step forward compared to Imagen 3, especially in text display — the element that has caused AI to “stumble” until now. However, the point worth discussing is that the market has changed. The “AI image generation” craze is no longer as new as it was when Midjourney or Dall-E first came out.

Some experts point out that users are gradually getting bored with AI-generated images, because most of these works lack soul, creativity is limited to commands and are often used for spam advertising purposes or as "just for show" illustrations on social networks.

In other words, Imagen 4 and Ultra, while technically more powerful, still fall short of the emotional and artistic breakthroughs that professional artists and designers can still do much better. That’s why Google needs to do more than just improve its prompt compliance if it wants to dominate the already saturated AI image creation market.

2. Google Gemini privacy changes

In addition to introducing Imagen 4, Google also caused a stir in the technology community when it sent an email announcing the new privacy policy of Gemini - an AI chatbot tool that is being oriented to replace Google Assistant in the near future.

In a recent email, Google said Gemini will soon allow users to use apps like Phone, Messages, WhatsApp, and Widgets on their devices, even when the “Gemini Apps Activity” feature is turned off. The change is expected to roll out starting July 7, prompting concerns that Gemini could “eavesdrop” or “read” personal content.

In fact, according to analysts, this change is not as new as many people fear. Before the new policy, users were required to enable “Gemini App Activity” to use utilities such as sending messages and making calls via Gemini. This meant that all chat content had to be stored on Google servers, and this data could be used to train AI models.

This rule has been in place for years, back when we used Bard, the predecessor to Gemini. Even when Gemini is turned off, Google still stores data for 72 hours for technical purposes. So privacy is already the price users have to pay for taking full advantage of the AI ​​features on Android phones.

What’s new this time is that Google has “relaxed” for users: you can still use features like calling, texting, setting alarms, and controlling widgets without turning on Gemini App Activity. When history is turned off, Google will not use the data to train AI, and it will not be stored for more than 3 days, unless the user turns this feature on themselves.

3. Google “clears the way” for Gemini to replace Google Assistant

Tech circles believe that this move paves the way for Gemini to become the “default virtual assistant” on Android, completely replacing Google Assistant in the near future. On Pixel phones, Gemini has officially taken on the role of the main virtual assistant, handling tasks such as making calls, sending messages, and controlling smart devices in the home.

In the latest announcement, Google emphasized:

“This update is great for users: they can now use Gemini to complete everyday tasks on their mobile devices like sending messages, initiating phone calls, and setting timers while Gemini Apps Activity is turned off. When Gemini Apps Activity is turned off, their Gemini conversations will not be reviewed or used to improve our AI models.”

This means users will have more control over their personal data, but it also raises the question: will the AI ​​experience be limited? Because without data sharing, the personalization ability that is AI's greatest strength may no longer be fully utilized.

From another perspective, Google's new policy is seen as a commitment to greater transparency about privacy, as AI companies face a barrage of criticism over the misuse of personal data to train models.

In fact, Google is not the only company under scrutiny. OpenAI - the company that developed ChatGPT or Meta with its own AI products - has also been sued many times for collecting data without users' knowledge. Allowing users to turn off history while still using Gemini's features is considered a smart move, alleviating users' concerns and avoiding legal troubles.

However, this also poses a challenge for Google: how to maintain the quality of AI models when the amount of training data is limited if users turn off storage en masse?

4. Gemini – An important link in Google's AI ecosystem

More than just an AI chatbot, Gemini is seen as an important link in Google's AI ambitions. This tool not only answers questions and creates content, but also connects closely with other Google services, from Gmail, YouTube, Google Photos to Drive.

With the new move, Google clearly wants Gemini to become a "super virtual assistant", a comprehensive AI platform, replacing Google Assistant and directly competing with Siri (Apple) or Alexa (Amazon).

Looking back, Imagen 4 and Imagen 4 Ultra show that Google still has a long way to go if it wants to surpass Midjourney or Dall-E in the field of AI image generation. Although the technology has improved, the “human touch” in photos is still something that AI cannot completely replace.

On the other hand, Gemini’s new privacy policy is considered a cautious but calculated move. It both soothes user concerns and prepares for Gemini’s ambition to become Google’s “right-hand man” on all mobile devices.

Will Google continue this momentum or will it soon be overtaken by its competitors? It remains to be seen, but one thing is certain: artificial intelligence is no longer a “distant future” but is present in every device we use every day, whether we are willing to share our data or not.