ChatGPT, the AI chatbot that can’t stop lying, has exploded in popularity over the past few weeks. Despite many problems with the technology, Google and Microsoft are pushing ahead with plans to turn web searches into conversations.
Microsoft has invested heavily into the company behind ChatGPT (and DALL-E), known as OpenAI, and has already been testing an upgraded version of the Bing search engine with the new technology. Instead of typing in traditional search queries and clicking through the results (or just reading the snippet provided by the search engine), screenshots provided to The Verge indicate Bing’s new mode will accept normal questions and print out conversational replies. Bing also provides citations for each statement, similar to Wikipedia, which is a notable problem with regular ChatGPT.
Microsoft is planning an event for tomorrow, which will likely be the full reveal of the new Bing features, but Google already has details to share about its take on the concept. Google officially announced “Bard” today, a conversational search tool based on the Language Model for Dialogue Applications (or LaMDA for short) technology that the company has been slowly developing for several years.
Google said in a blog post, “Bard seeks to combine the breadth of the world’s knowledge with the power, intelligence and creativity of our large language models. It draws on information from the web to provide fresh, high-quality responses. Bard can be an outlet for creativity, and a launchpad for curiosity, helping you to explain new discoveries from NASA’s James Webb Space Telescope to a 9-year-old, or learn more about the best strikers in football right now, and then get drills to build your skills.”
The AI features won’t just be limited to Bard, though. Google plans to roll out its explanation features in normal search queries, which will attempt to “distill complex information and multiple perspectives into easy-to-digest formats.” The company says those upgrades will roll out “soon.”
Even though Google is promising it will monitor responses for quality and safety, it’s hard to be optimistic about conversational technologies like ChatGPT for informational purposes. Finding false information is already easy with traditional search engines, but AI models like ChatGPT can easily misinterpret data or even make up information entirely. The leaked Bing screenshots revealed inline citations, but there doesn’t appear to be anything like that in the screenshots provided by Google.
At a time when misinformation is a growing problem online, it’s frustrating to see companies like OpenAI, Google, and Bing pushing ahead in a race to be the first widespread AI search tool before they are robust enough. Google’s model might be more accurate — it’s not directly based on ChatGPT, like Bing’s feature is expected to be — but there’s nothing available right now that can actually think for you.
