- Users have the option to explore an earlier iteration of Google’s GPT-4 competitor Gemini via Bard.
- Despite upgrades, some individuals have pointed out that the chatbot still struggles with basic factual accuracy.
- Bard seems hesitant to engage in discussions on controversial topics, often redirecting users to Facebook instead.
Google has recently introduced Gemini, its competitor to GPT-4, but it seems no closer to resolving its AI hallucination problem.
Critics on social media have called out Google’s Bard AI, which currently operates a limited version of its GPT-4 counterpart Gemini, for persisting in providing inaccurate responses to simple queries and avoiding sensitive topics by suggesting users search on Google instead.
Reports first surfaced on Techcrunch highlighting Bard’s inability to comprehend requests for translations and its lack of knowledge about the winner of the Best Actor award at the 2023 Oscars.
When asked about the Emmy nominations, Bard mistakenly listed Squid Game and Stranger Points as contenders for Outstanding Drama Series, despite the correct shows being Squid Game and Stranger Things, which aired their latest seasons in 2021.
Users on platform X have also noticed Bard’s reluctance to address ongoing contentious issues such as the conflict in Gaza, opting to advise users to utilize Google search for the latest updates. In comparison, competing AI models like Elon Musk’s Grok, Bing, and ChatGPT Pro offer more comprehensive responses and access to real-time information.
Users of Bard are cautioned to verify the information provided, as Google acknowledges the potential for misleading responses, a common trait among conceptual AI chatbots.
Gemini was unveiled on Wednesday after much anticipation, with Google positioning the AI model as a significant milestone in catching up to competitors like OpenAI.
Google was reportedly caught off guard by the enthusiastic reception of OpenAI’s ChatGPT robot upon its launch last month.
In response, Google introduced its Bard bot in February to address the gap. However, concerns arose when advertisements for the ChatGPT competitor revealed inaccuracies in its responses, leading to a sense of embarrassment for the tech giant.
While Bard currently operates on Gemini Pro, Google is gradually introducing Gemini to the market. The upcoming version, Ultra, is set for an early release next year.
Google claims that Ultra surpasses OpenAI’s GPT-4 model in performance, although the Pro version falls slightly short of that benchmark, outperforming the PPT-3.5 model that powers the free ChatGPT variant.
Questions have been raised about the capabilities of Google’s latest AI offering. A video showcasing Gemini effortlessly handling natural language and complex voice commands amazed viewers online, but a Google spokesperson later admitted to Bloomberg that the video had been edited to enhance its impact, prompting internal scrutiny within the company.
A request for comment from Google by Business Insider outside regular business hours did not receive an immediate response.
