Google begins to experiment with podcasts generated by AI in searches results

Google, as announced in its official blog, has begun to try a new experimental function in its search engine Audio Overviews. It is a tool that generates audio summaries with AI on the results of a consultation, presented in a conversational format that simulates a podcast with two presenters. It is a function that we already met in the guide on notebooklm and that could now be integrated into the results of the searches, integrating with the summaries generated by AI.

This functionality, driven by the company’s gemini models, seeks to offer users a form of Absorb information without even reading a text. Without a doubt, it seems ideal for those who perform several tasks at the same time or simply prefer the audio format. The option, for now, is Available in Search Labs For users in the United States and only generates summaries in English.

When activating it, in certain searches an option will appear to generate the audio. The result is a clip of between 3 and 5 minutes in which two voices generated by the subject. As reported on PC Gamer, although voices are very realistic And they use a conversational language, interaction is not entirely natural, giving the feeling that presenters read a script. Something interesting is that, to guarantee the veracity of the facts presented in the podcast, Google shows the pages used to create the summary directly in the player.

While the tool can be useful for people with visual disabilities or to consume content passively, it also raises concerns. Without going any further, the realism of the voices could lead to some users to assume that they are real people, which increases the risk of spreading misinformation generated by AI. It also discourages that the user clicks on the links to check the facts and generate the traffic that the means to finance so much need.