Google Gemini live demo failed twice during the Pixel 9 event ...
Google launched the Pixel 9 series of devices at the Made by Google event. Before introducing the new Pixel 9 Pro, Pixel 9, and Pixel 9 Pro Fold, Google showcased the new Gemini features through a live demo. Unfortunately, the demo failed not once, but twice, while trying to demonstrate how Gemini's mobile AI features integrate with the calendar app by providing information through an image.
Gemini Fails to Respond
Upon submitting the prompt via voice command, Gemini failed to respond both times. Surprisingly, the demo only worked when the device was switched to a different Samsung Galaxy S24 Ultra. Google's attempt to showcase the new Magic Editor feature, as an expansion on the Magic Eraser, also faced challenges. The AI created an awkward object in the second photo after trying to manipulate an image by inserting a hot air balloon. This incident highlights the limitations of generative artificial intelligence in its current state.
Recognition of Faults by Google
Google acknowledges that Gemini is a creativity and productivity tool that may not always be reliable, particularly with tasks like generating images or text related to current events. The concept of "hallucinations" in Large Language Models contributes to instances where AI may make mistakes. The technology is still evolving rapidly, and even major AI players like OpenAI recognize that their models may occasionally provide inaccurate information.
In light of past mishaps, Google had to disable features in the Gemini app temporarily to prevent the generation of inaccurate or offensive images. It is evident that the road to achieving Artificial General Intelligence (AGI) is a challenging one, with unexpected hurdles along the way.
Conclusion
Despite the setbacks faced during the live demo at the Pixel 9 event, Google continues to push the boundaries of AI technology. The journey towards more reliable and trustworthy AI systems is ongoing, with valuable lessons learned along the way.
