Exploring Bard: A Review of Google's Conversational AI

As technology continues to advance, so does the field of artificial intelligence. One notable development in this realm is Bard, an experiment by Google based on the same technology behind LaMDA (Language Model for Dialogue Applications). As a conversational AI model, Bard aims to be a creative and helpful collaborator, supercharging users' imagination, boosting productivity, and bringing ideas to life. In this blog post, we will delve into the unique features and performance of Bard based on a compilation of reviews from college students and my own personal experience.

Collaborative Generation and Drafting:

Bard stands out with its innovative approach to collaboration. When interacting with the AI, users are presented with three different drafted answers per prompt. This feature allows for exploration and choice, providing a diverse range of responses. However, it's worth noting that once the conversation progresses, the previous drafts are no longer visible.

Limitations in Mathematical Abilities:

One area where Bard, along with other language models, exhibits limitations is in mathematical calculations. For instance, when asked to determine the character count of a sentence, Bard fell short by providing an incorrect response. While it acknowledged the correct answer upon correction, the initial inaccuracy and subsequent iterations raised concerns about its mathematical capabilities.

Creativity and Storytelling:

Bard's potential for creativity shines through in its storytelling abilities. When prompted to write a story about a black cat named Cocoa who lives in an office, Bard produced an incredibly adorable narrative. Its writing style was reminiscent of a storybook, capturing the essence of the character and the setting. However, personal preferences may vary, and some users may prefer the more cohesive and mature tone offered by ChatGPT in its storytelling.

Staying Up-to-Date:

Bard's ability to provide real-time information and news updates falls short compared to its search engine counterparts. While Bing was able to deliver curated news about startup activities in Malaysia, ChatGPT acknowledged its limitations and recommended external resources for the latest startup news. Bard, on the other hand, admitted its inability to assist in this regard.

Conversational Style:

The effectiveness of Bard in emulating a close friend from Malaysia who is sarcastic yet kind varies. Bing demonstrated its ability to deliver sarcasm with witty responses, using emojis to add a personal touch. ChatGPT impressed with its consistent tone and ability to remember the initial objective throughout the conversation. However, Bard's performance in this area was found to be subpar, lacking the intended sarcasm or humor.

Conclusion:

In evaluating the performance of Bard, it becomes evident that the AI model still has room for improvement. While it offers unique collaborative drafting and showcases creativity in storytelling, Bard's limitations in mathematical calculations and real-time information retrieval raise concerns. Comparatively, ChatGPT displayed commendable consistency, and Bing excelled in providing up-to-date news. As an experimental project, Bard may undergo further refinements in the future. For now, users may find other AI models better suited to their specific needs and expectations.