Gemini Live Is Far From Perfect And Fails To Live Up To Its Hype, Users Claim

gemini-live-is-far-from-perfect-and-fails-to-live-up-to-its-hype,-users-claim
Gemini Live Is Far From Perfect And Fails To Live Up To Its Hype, Users Claim

Google has started rolling out Gemini Live as one of its biggest and most sought-after AI offerings.

It was described as a chatbot that gave users the closest experience to chatting with a human-like assistant. The issue that people are facing is the bland personality, not to mention the unreliability factor attached.

Some are linked to OpenAI’s Advanced Voice Mode. It was outlined to be the most realistic voice and also had the ability for users to interrupt the chatbot when it was delivering a response.

The issues many are having with the launch are its inability to solve complex issues and the fact that it fails to give a natural tone during the engagement. Yes, it’s better than Google’s previous versions but is it really enough?

The chatbot is yet to address the growing number of problems that users have discussed. Inconsistencies are on the rise as are hallucinations. Some feel it’s a fancy version of the older attempts. We’re sure Google might not be too happy to hear that.

The tone is outlined to lack passion. You cannot make adjustments to what you’re hearing including the pitch or the speed at which it’s jetting out responses. Clearly, this makes it incomparable to OpenAI’s Advanced Voice Mode which offers just that. However, the chatbot is more polite and does not hesitate when speaking. It can similarly handle several chats taking place at the same time.

Another feature described by Google was the ability to assist users with job interviews. When one user fed it with information that was ‘out of line’, the chatbot appreciated the replies and implied that they were on the right track. Giving false hopes is certainly not appreciated, many wonder.

Gemini Live was also said to display strange actions. It would recall things discussed from previous chats and try to link them to current ones. This is the problem of hallucinations that we discussed above. Yikes, Google needs to work on this one as that can be a huge turn-off for users.

Outdated information wasn’t caught by the chatbot and users had to remind it of the experience. On that note, the chatbot behaved politely and thanked users for finding the mistake of outdated data.

There were some instances where Gemini Live was so adamant about being right that it would make anything wrong seem correct. Imagine asking the chatbot to play a game and the move it makes is so convincing that it makes you wonder where you went wrong. Google has a lot of work to do. What do you think?

“To me it doesn’t feel conversational. More like an AI that is ok if you interrupt and fast to recover from interruptions. With respect to integration to Workspace – I definitely see the flopsweat. It’s super slow to roll out. Rather than one or two highly compelling features there is a lot of spaghetti throwing to see what sticks to the wall. I’m pretty sure Google was all out panicked by openai. IMO the Google offering is a lot of trying to catch up.”, noted a user on Reddit.

In general, the responses were so wishy-washy in nature. Some replies the chatbot refused to touch base on. For instance, politics and the current war on Gaza had some bizarre narratives. It even warned against discussing certain topics. Let’s not even get started on the technical issues.

Read next: Zoom Announces Revolutionary Single-Use Webinar Offering That Can Host 1 Million Attendees