Artificial intelligence enthusiasts will have to wait a little longer to witness the debut of Gemini, Google’s next language model intended to compete with GPT-4, the ChatGPT engine developed by Open AI. The Mountain View-based company has decided to postpone the event that was scheduled for this week where this innovative technology was going to be presented. Curiously, this week it was leaked that Open AI has also postponed the launch date of its GPT Store after the commotion caused by the dismissal and subsequent reinstatement of its CEO and founder, Sam Altman.
Google intended to present Gemini in Washington, New York and California, The Information publication has announced. However, the corporation has chosen to reschedule these events to early 2024, in order to address certain technical issues that have arisen in the process.
Until now, there was no knowledge of Google’s plans to introduce its latest creation in artificial intelligence during this week, but according to The Information, these events were intended to showcase the capabilities and characteristics of Gemini before American politicians and legislators.
This strategy is no coincidence, especially considering that at the end of October, President Joe Biden issued an executive order to ensure the “safe and reliable development and use” of artificial intelligence. This order calls on companies that develop AI models to be more transparent about how they work, establishing security guidelines, such as the use of watermarks and similar methods to identify synthetically generated content.
However, it seems that Google needs more time to perfect Gemini. The company aims for this new model to be positioned as an alternative that matches or exceeds the capabilities of GPT-4, with a key focus on native support for multiple languages. But this is where the main problem arises. Gemini apparently fails to correctly process commands entered in languages other than English.
As for the rest of the Gemini’s capabilities, Google has revealed few details so far. Announced at the I/O 2023 event, run by Google DeepMind, it is described as a next-generation multimodal AI model. However, the information available is limited, and there is speculation about its integration with current Google products and services.
It is striking that Google, which already has PaLM 2, also presented at I/O 2023, is developing another model like Gemini. In fact, PaLM 2 powers Bard, the chatbot that competes with ChatGPT, and has found applications in fields such as medicine through Med-PaLM 2.
It will be interesting to see how Google addresses and solves technical challenges, and how it plans to integrate Gemini into its current complex ecosystem. At a time when attention is increasing on training methods and safeguards in artificial intelligence, Google wants to ensure that Gemini meets the highest standards of transparency and functionality.