Hallucinations - an industry term used to describe the situation where an AI platform generates responses that are inaccurate or don't make sense but are presented as factual - is an area where platforms and the public may not see the same thing. * Only 5% of user interactions...
We gathered 4,200 responses from the four leading AI platforms and had 400+ users pick which ones they preferred. The emerging winner so far is GPT-4, but it’s a close race. GPT-4 leads the way, with responses preferred 55% of the time. That's a significant advantage...
While GPT-4 does very well overall in the comparison rating, there was a consensus among those who DIDN'T prefer GPT-4 responses as to why: GPT-4 is not preferred when there is “TMI” or “too much information.” Implication Striking the balance of “just enough” information - but not too...
In blinded side-by-side testing, users select their preferred AI response to a series of queries and describe why. As a result, we see how the four platforms compare across hundreds of users and various question types. * Bard stood out for responses that were viewed as most complete, accurate, and relevant....
Across all users, the top three reasons for choosing a preferred AI response in our blinded tests are: 1. Feels accurate 2. Feels relevant 3. Feels complete In addition, among younger users, the attribute “feels human like” increases in importance. This video of a young user asking Bard for a...
In our testing, the top three types of question users submitted to the AI platforms were in the areas of: 1. Learning & Education 2. Technology 3. Media & Entertainment Looking at differences by generation, we find: * Millennials more interested in Money, Finance, Insurance * Gen Z more interested in Media,...