Bard vs Chat GPT Writing Science Fiction with AI
People & Blogs
Introduction
Recently, I received access to Google Bard, Google’s equivalent of Chat GPT. Intrigued by the comparison, I decided to put both AI systems head to head by having them generate a 3,000-word science fiction story. The results were unexpectedly thought-provoking.
Initial Impressions
For context, I’ve been using Chat GPT for a while to write fiction, and the results have been reasonably satisfying. On the other hand, my initial experience with Bard over a week was somewhat disappointing. Bard often claimed it couldn’t perform certain tasks and, when it did, the results were underwhelming.
Story Concepts
Chat GPT's Story Concept
I started by asking Chat GPT to provide story suggestions based on themes around contact with alien life and the future of humanity. Chat GPT gave insightful suggestions, including Arthur C. Clarke's "2001: A Space Odyssey," "Alien," "The Expanse," and works by E.M. Banks.
I then asked Chat GPT to outline a short science fiction story with a twist ending. It suggested, "The Language of the Stars," which involves a team receiving an alien message and trying to decode it. Though the concept was intriguing, the twist was a bit predictable.
Bard's Story Concept
For Bard, I asked the same initial questions. After suggesting several stories, it provided a compelling story outline where aliens are astronauts from the future. While this concept appealed to me, Bard's execution was not as robust as Chat GPT's.
Writing the Stories
Chat GPT's Execution
When asked to write a 3,000-word science fiction story, Chat GPT notified me that it couldn’t generate the full story in one response. Instead, it offered to break the story into sections, with rich text and suspense.
Bard's Execution
Bard, however, did not retain previous conversations, forcing me to re-enter questions. This time, it returned a different story outline that didn’t match the initial concept I liked. When asked to write the story, the output was far shorter and less detailed than Chat GPT’s, more akin to an outline than a fleshed-out narrative.
Creating AI Art
I also explored generating AI art based on the story descriptions using a tool called Prompt Mania. Descriptions from both Bard and Chat GPT were used to create visual prompts for platforms like MidJourney. While both sets of prompts were useful, Chat GPT’s descriptive detail provided richer imagery.
Conclusion
In summary, Chat GPT outperformed Google Bard in terms of quality and consistency. While Bard showed potential with interesting story concepts, its execution in writing detailed prose was lacking. This may pose a concern for Google as the future of AI potentially shifts from search engines to AI-driven content creation.
Keywords
- Google Bard
- Chat GPT
- AI writing
- Science fiction
- Story generation
- Alien contact
- Future of humanity
- AI art
FAQs
Q: Which AI system produced better story concepts?
A: Bard suggested an intriguing concept where aliens were astronauts from the future. However, Chat GPT’s execution in prose was superior.
Q: Can Bard remember previous conversations?
A: No, Bard did not retain previous conversations, unlike Chat GPT which maintains a chat history.
Q: What tool was used for generating AI art based on the stories?
A: A tool called Prompt Mania was used to generate detailed AI art prompts.
Q: How did the text generation capabilities of Chat GPT compare to Bard?
A: Chat GPT was more detailed and coherent in generating lengthy prose, whereas Bard’s output was more outline-like and less immersive.
Q: Did the experiment involve other AI models?
A: No, the experiment was mainly between Google Bard and Chat GPT. Tools like MidJourney and Prompt Mania were used for supplementary AI art creation.
Q: Why was the twist in Chat GPT's story considered predictable?
A: Despite being well-written, the twist could be anticipated by readers familiar with common science fiction tropes.