Since its launch a few days ago, Gemini, Google’s latest artificial intelligence (AI) model, has generated significant excitement. However, European users will have to wait before they can fully enjoy the features offered by this revolutionary technology. Nevertheless, there are already ways to bypass this geographical limitation by using a VPN to access Gemini from France or other European countries.
To use Gemini in Europe, simply install a VPN software and select a server located in the United States where this AI is currently accessible. However, it is important to remain cautious and not violate the service’s terms of use.
The first impressions of AI Gemini
To better understand the potential of Gemini, we chose to connect from an American IP address using a VPN. The resulting chatbot explains that it uses the Gemini Pro model developed by Google. After importing a screenshot showing a rabbit jumping rope, we asked Gemini what sport this adorable little animal was practicing. The provided answers were coherent and relevant, demonstrating significant advances in the natural language processing capabilities offered by this new generation of AI.
An overview of Gemini’s capabilities through an impressive video
Recently, an impressive video has been circulating on social media to demonstrate Gemini’s capabilities.
Lasting six minutes, this video shows a smooth and natural conversation between a user and the artificial intelligence. Gemini is capable of recognizing a hand-drawn duck, locating a paper ball hidden under a cup, and suggesting knitting items made from two different colored balls of wool.
Factual errors in the presented video
However, it should be noted that the video has been edited to give the impression of a more natural conversation. In the description posted on YouTube, Google admits that “latency has been reduced and Gemini’s responses have been shortened for conciseness.” Additionally, some elements are added to clarify that the chatbot must identify a game from three images of hands or categorize three images representing the sun, Earth, and Saturn according to their distance from the Sun.
Oriol Vinyals, Vice President of Research at DeepMind (Google’s UK subsidiary specializing in AI), thus confirms that “all commands and results in the video are real but shortened for conciseness.”
Mixed initial feedback on the use of Gemini
Beyond this video, the initial steps of Gemini have left some people skeptical. The first testers who integrated it into Bard – Google’s chatbot – have mixed opinions. Indeed, it seems that this version of Gemini is not the most powerful model developed by Google.
Nevertheless, it is important to highlight that this feedback concerns an early use of Gemini. As researchers and developers continue to work on this project, it is likely that its capabilities and usefulness will improve significantly.
The future of Gemini and conversational AIs
The technological advances made with Gemini show real potential for future generations of artificial intelligences. While current models like OpenAI’s GPT-4 or Gemini Pro are already experiencing some success, there is still a lot of work to be done to achieve a completely smooth and natural human-machine interaction.
In summary, although Gemini is still in the development phase and its availability in Europe is limited, this new artificial intelligence model from Google promises significant advancements in the field of human-machine communication and could revolutionize the way we use information technologies in the future.
