Gemini is an increasingly good chatbot, but it’s still a bad assistant – Ars Technica

🔍 Summary:

Google is transitioning from its Google Assistant to a new virtual assistant named Gemini, based on generative AI, with plans to complete the switch by the end of 2025. Despite the advanced capabilities promised by generative AI, such as integrating data across multiple apps, the technology is proving problematic. The fundamental issue with Gemini, and generative AI in general, is its tendency to “hallucinate” or produce plausible but incorrect information. This flaw has led to errors like generating fake tracking numbers and scheduling errors, which can be frustrating and time-consuming for users.

The unreliability of Gemini contrasts with the more straightforward functionality of Google Assistant, which, despite its limitations and bugs, generally handles simple tasks like setting timers and sending messages reliably. The transition to Gemini also means developers must rebuild tools and integrations from scratch, adding to the challenges posed by this new technology.

As Google pushes forward with Gemini, aiming to keep up with competitors like OpenAI, the tech community remains skeptical about its readiness to handle important tasks reliably. The upcoming Google I/O event will likely focus heavily on Gemini, where Google will need to demonstrate significant improvements to gain trust from users and developers alike.

📌 Source: https://arstechnica.com/gadgets/2025/04/gemini-is-an-increasingly-good-chatbot-but-its-still-a-bad-assistant/