Controversy surrounds Google’s recent launch of its much-anticipated Artificial Intelligence model, Gemini. A six-minute promotional video, meant to highlight Gemini’s advanced capabilities, has come under scrutiny for not being conducted in real-time. This controversy adds to the company’s history of questionable demonstrations earlier in the year, raising concerns about the transparency of Google’s AI showcases.
The Gemini Launch and Demonstrative Video
In a bold move just weeks before the year’s end, Google introduced Gemini, touting it as its largest and most capable AI model. To accompany the launch, the company presented a six-minute video showcasing Gemini’s prowess. The video featured spoken interactions between a user and a Gemini-powered chatbot, demonstrating the AI’s ability to recognize visual images and physical objects with impressive accuracy.
However, the excitement surrounding Gemini quickly turned into controversy as users and media outlets discovered that the video was not a real-time demonstration. Google later confirmed to Bloomberg that the demo consists of still images and text prompts, deviating significantly from the impression the video initially conveyed. The company’s disclaimer on YouTube mentioned reduced latency and shortened Gemini outputs but failed to make this clear within the video itself.
Amid mounting criticism, Google responded to inquiries from CNBC, stating that the video was an “illustrative depiction of the possibilities of interacting with Gemini” based on real multimodal prompts and outputs from testing. The company emphasized that Gemini’s capabilities were genuine but acknowledged the use of edited content for brevity in the promotional video.
Revisiting Demos Before Gemini
This controversy is not the first time Google has faced backlash for its AI demonstrations. Earlier in the year, the company’s employees labeled its AI chatbot demo as “rushed” and “botched.” This prior incident coincided with Microsoft’s plans to showcase Bing integration with ChatGPT. The repetition of scrutiny raises questions about the thoroughness and accuracy of Google’s AI showcases.
Gemini in Competition with OpenAI’s GPT-4
Google’s Gemini is positioned in direct competition with Microsoft-backed OpenAI’s GPT-4, which has been hailed as the most advanced and successful model to date. In an attempt to establish Gemini’s superiority, Google released a white paper claiming that Gemini’s most powerful model, “Ultra,” outperformed GPT-4 across several benchmarks, though only by incremental margins.
Despite the controversy, Google remains optimistic about Gemini’s future. The company has announced that access to Gemini Pro will open on December 13, expressing eagerness to witness the creative applications users will develop. As the AI landscape becomes increasingly competitive, the controversy surrounding Gemini’s demo raises broader questions about the industry’s commitment to transparent and accurate demonstrations.