The metaverse dominated the previous year, but the tech giants are already shifting their focus to something else. In 2023, AI chatbots will be popular, but can you trust them? Within a few weeks of Microsoft receiving positive news for the announcement of its ChatGPT-integrated Bing revamp, Google has launched Bard, its own chatbot, for public testing. These AIs have already demonstrated that they are prone to making careless errors, but the current uproar is on a whole new level. Now that Bard and Bing are mentioning one another, the false information is being repeated until it is only a memory.
This joke may have caught your attention while it was being circulated online. With Google’s propensity to discontinue services that are not immediate, resounding successes, someone on Twitter questioned Bard about when the company will shut it down. In response, Bard claimed that it had already been shut down, citing a single humorous Hacker News comment. Bard said it (Bard) had already been shut down after misinterpreting the humour. Funny, yes?
This comedy of mistakes is still going on. The Verge claims that Microsoft’s OpenAI-powered Bing capitalized on that error. When questioned if Bard had been shut down, Bing cited a report from Windows Central about Bard’s awkward response. According to the story, which Bing misinterpreted, it was said that Bard had already passed away. It’s only been a few weeks since the release of these ostensibly groundbreaking chatbots, and now they seem to be eating themselves.
Although Bing has now been modified to appropriately respond that Bard is still alive, this exchange highlights an important issue for AI: It is unsure of what is real. Humans aren’t perfect in that aspect, after all. Who among us hasn’t fallen for an April Fool’s joke online and forgotten what day it was? These algorithms lack humour and look absurdly credulous, often accepting a single comment or news piece without understanding its greater context. Perhaps Google and Microsoft should set their bots to reject content uploaded on April 1st in order to get an early start on this.
Fast-moving businesses are integrating these tools into their goods. The GPT bot was included to Microsoft’s Bing search engine. Bing is still a key product for Microsoft even though it does not compete with Google. Google, whose Workspace products are used by corporations and educational institutions all throughout the world, has announced intentions to incorporate AI akin to Bard. We may look back on this time period in a few years as an important lesson on the limitations of AI.
Remember that OpenAI was previously rather wary of making its GPT models available to the general public due to their propensity to disseminate false information. Now it seems as though those worries were justified. Is referring to them as “experiments” or “beta” a sufficient defence for the crude sourcing practices?