Microsoft's Bing+ChatGPT3 integration led to an accelerated launch of Bard by Google this week (both are in limited, 'soft' launch at present).

Some are predicting the end of search, foreseeing that consumers will quickly switch their search experience preference away from pages of search result listings to a single answer provided in real time by generative AI.

A key risk with generative AI technologies is the high degree of inaccuracies and, not so occasionally, sheer fabulism present in the answers provided.

You need only look at Google's advertisement promoting Bard for an example of this: it gave a believable, but very inaccurate, response to a query about the James Webb Space Telescope.

One of the hardest elements of digital literacy education is teaching people how to find and assess information for accuracy and reliability. Students are taught to prefer known, reputable content sources, to verify accuracy by comparing multiple sources, and to use critical thinking skills rather than merely accept published writings.

The single generated answer - without sources - presented by tools like ChatGPT3 and others, will make it harder for consumers to determine if the answer is accurate and credible.

Moves by startup search companies like Neeva to incorporate sourcing and citations in live, generated answers is a good first step.

But we also need to recognise that many (most?) people won't use these details to do their own validation.

It should be possible to take it a step further, and add an additional content analysis step (or steps) that allows an accuracy or confidence score to be made visible to the user, to assist their choice of how and when to use generated answers.

#google #microsoft #experience #ai #generativeai #digital #education #digitalliteracy #digitallife #content #chatgpt3 #chatgpt #bard #startup