Lo scorso anno Google aveva reso disponibile, ad un numero ristretto di utenti,  NotebookLM, uno strumento di intelligenza artificiale che nasceva per aiutare a prendere appunti e studiarli meglio.

A  giugno 2024 è stata estesa la fruizione in oltre 200 paesi di tutto il mondo.

Oggi, NotebookLM, offre funzionalità avanzate come la generazione automatica di riepiloghi, una guida allo studio, suggerimenti di domande frequenti sui contenuti dei documenti caricati e la possibilità di generare un podcast colloquiale. NotebookLM consente di creare “block notes”, delle vere e proprie aree di lavoro, all’interno delle quali è possibile caricare fino a 50 documenti su cui far lavorare l’IA ed interagire con essa. A differenza di chatbot generici, opera esclusivamente sui documenti inseriti, creando un’esperienza personalizzata.

Questa applicazione utilizza il modello di linguaggio avanzato Gemini Pro, che migliora la comprensione e il ragionamento sui documenti caricati.

Gli utenti possono trasformare le loro note in progetti strutturati, come documenti propedeutici a briefing o bozze di piani di marketing. NotebookLM può suggerire azioni in base al contesto, ad esempio riassumendo una parte selezionata di un testo o affinando la prosa dell’utente durante la stesura di una nota.

Integra un chatbot al quale è possibile formulare istanze in forma di prompt (alcune di queste vengono automaticamente proposte all’utente) e le risposte di cui si intende mantenere evidenza possono essere salvate come “Note” nell’area di lavoro (il “block note”).

Ma il vero punto di svolta è la recente introduzione della funzionalità Audio Overview.

Questa funzione trasforma i documenti caricati in dialoghi simili a podcast, con due voci generate dall’ Intelligenza artificiale, una femminile una maschile, che discutono in modo realistico e raccontano e spiegano in modo informale il testo contenuto nel documento originale.

È sorprendente come le due voci non solo sembrino naturali, ma interagiscano con fluidità: si alternano al momento giusto, sottolineano con enfasi le informazioni l’una dell’altra e aggiungono piccole considerazioni che, anche se generate da un’IA, arricchiscono il testo e rendono più facile capire anche argomenti complessi o noiosi. Tuttavia, nonostante i documenti caricati siano in lingua italiana, l’audio generato è disponibile, per il momento, solo in lingua inglese.

Con Notebook LM l’impegno di Google nell’ambito educativo viene ulteriormente potenziato dall’utilizzo dell’intelligenza artificiale, proseguendo l’impegno nel facilitare e guidare l’apprendimento già da tempo avviato, ad esempio attraverso la sezione “Google for Education”, in cui è possibile visualizzare i diversi strumenti messi a disposizione per la didattica, come Google Classroom, che è molto usato in Italia.

Impegno che è ulteriormente emerso durante lo stesso evento Google I/O della scorsa primavera in cui si presentava NotebookLM quando, Ben Gomes, Senior Vice President di Education & Learning di Google, ha presentato anche Illuminate, uno strumento sperimentale sempre basato sull’IA che, come Audio Overview genera una conversazione a due voci che discutono i punti principali di articoli accademici e di ricerca.

Come funziona NotebookLM?

Per iniziare ad utilizzare questo innovativo strumento, bisogna visitare il sito notebooklm.google e selezionare l’opzione “Prova NotebookLM”. Dopo aver effettuato l’accesso, apparirà l’interfaccia di NotebookLM e, come primo passo, bisogna creare un nuovo notebook che,come abbiamo detto, rappresenta un’area di lavoro dedicata.

Successivamente è necessario caricare, all’interno di quest’area di lavoro, le fonti desiderate, che possono essere in formato Google Docs, PDF, URL o anche testo copiato.

Dopo l’upload delle fonti, NotebookLM genera rapidamente una guida al notebook, uno strumento utile per comprendere ed esplorare le fonticaricate.

Questa guida include:

Riepilogo: l’IA analizza le fonti e genera un breve riepilogo del contenuto, offrendo una breve panoramica delle informazioni principali

Domande suggerite: Alcune domande per iniziare ad esplorare il materiale.

Guide preformattate: guide più dettagliate in vari formati.

I formati proposti sono:

• Domande Frequenti (FAQ): L’IA analizza il materiale caricato e suggerisce automaticamente domande pertinenti, aiutando l’utente a esplorare meglio il contenuto.
• Guida allo Studio: Fornisce suggerimenti per approfondire e comprendere meglio il materiale caricato, creando un percorso di apprendimento personalizzato.
• Sommari: L’IA crea un riepilogo dei punti principali del documento, consentendo una rapida panoramica delle informazioni essenziali.
• Sequenza Temporale: Organizza gli eventi o le informazioni presenti nei documenti in ordine cronologico, l’analisi del contenuto in forma temporale.
• Documenti di Briefing: Sintetizza i dati in un formato strutturato, ideale per presentazioni o analisi rapide, rendendo più semplice la preparazione di riunioni o discussioni​

In basso a sinistra è possibile visionare il riepilogo, mentre in alto a destra si trova la sezione dedicata ad Audio Overview. Per ottenere il nostro podcast basterà cliccare su “Generate” e, in pochi minuti, viene creato un podcast interattivo che riassume e discute il contenuto caricato, lingua inglese.  È naturalmente possibile scaricare il file audio e ascoltarlo in un secondo momento proprio come accade con un podcast.

L’Audio Overview di quest’articolo con NotebookLM (inglese) e trascrizione

Speaker 1: Ever feel like the way we learn is like, right on the verge of something totally new?

Speaker 2: You sent over this article about Notebook LM, this AI from Google, and I gotta say, it’s got me thinking maybe AI is the big change, especially since it’s not just some exclusive thing anymore, right? It just launched in over 200 countries. What’s got you so curious about Notebook LM?

Speaker 1: Well, I think the thing about Notebook LM is, you know, it’s not just another app, you know, like for taking notes. It’s not just a chatbot either, spitting out the same old stuff. Imagine, like, this superpower digital notebook that’s actually trying to understand what you’re working on. So I upload my research papers, meeting notes, even those random thoughts I jot down—exactly, even those—and the AI actually digs in and makes sense of it all. It’s like having, like, an AI study buddy who can actually pull out the key takeaways from, like, a dense research paper, or even take my jumbled notes and, like, turn them into a study guide. That’s always been the dream.

Speaker 2: And it gets even more interesting. See, the engine behind Notebook LM, the thing that makes it tick, is Google’s Gemini Pro. That’s a super advanced language model. But the big difference, and this is key, is that it’s all about deep understanding. It’s designed to really get what you’re feeding it.

Speaker 1: Okay, now you’ve got my attention. Deep understanding is a whole other level. So how does this actually work? Like, if I gave Notebook LM a document, what would I actually see happen?

Speaker 2: So say you’re working on this big project, right? You’ve got a ton of research. You feed a document to Notebook LM, and boom, it gives you this concise summary. Right away, no more spending hours trying to find the important stuff.

Speaker 1: It’s like having that one friend, right? The one who reads everything and can tell you exactly what matters. We all need a friend like that.

Speaker 2: But it sounds like it does a lot more than just summarize.

Speaker 1: You got it, Notebook LM can also look at that document and come up with questions about it.

Speaker 2: Now, that’s smart. So instead of just reading and trying to remember, Notebook LM is pushing you to actually think critically about what you’re learning.

Speaker 1: Exactly. It’s like having a built-in study group right there in the software. So it’s not about making information easier to digest, it’s about making it easier to actually learn and use.

Speaker 2: 100%, and that’s where it gets really exciting, because it’s not about speed. It’s about changing how we use information to learn. It’s not about making things faster, it’s about making them better.

Speaker 1: But let’s get into some specifics. The article mentioned it can create all kinds of different guides. What were those?

Speaker 2: This is where things get really interesting for me. Like, imagine you need to learn about something new super fast. The article mentioned that Notebook LM can actually create different types of guides.

Speaker 1: What kind of guides are we talking about?

Speaker 2: So let’s say you’re going through this really complex research paper, and you want to be ready for any questions. Notebook LM can take that paper and make a list of FAQs, you know, like frequently asked questions. It’s a lifesaver if you have a presentation or something.

Speaker 1: Oh, that’s brilliant. So it’s not just summarizing. It’s like it knows what you might need to know next.

Speaker 2: Exactly, and when you’re trying to wrap your head around something historical, something with a lot of context? Right?

Speaker 1: Context is key for stuff like that. Notebook LM can create timelines—so helpful for any kind of historical event, scientific discovery, or even just seeing how an idea changes over time.

Speaker 2: I love that. It’s like they’re thinking about different ways people learn. Some people are visual learners, yeah, right? So timelines would be huge for them.

Speaker 1: What about the people who need things super organized, you know, like they’re getting ready for a big meeting? They need everything just so.

Speaker 2: Exactly. So for those folks, Notebook LM can even make what they called a briefing document, so it takes all the key points from your documents and puts them in this really clear, concise way.

Speaker 1: Okay, so that’s seriously impressive. It’s like having a personal assistant who’s also a research whiz.

Speaker 2: But the thing that really blew my mind, and you even circled this in the article, was this audio overview. How does it turn documents into like a podcast conversation?

Speaker 1: It’s wild, right? You basically upload your document and Notebook LM goes and creates this whole audio discussion, and get this—it’s between these two different AI voices, one male, one female, and they’re not just reading the document, they’re going back and forth, just like we are.

Speaker 2: That’s amazing. I get AI summarizing text, but a whole conversation—that feels next level. Did it sound natural?

Speaker 1: It’s actually pretty cool how they interact. Like, they’ll highlight each other’s points, offer a different perspective, even throw in little insights here and there, and it’s all in this conversational tone. It’s not perfect, you know, but it’s definitely engaging.

Speaker 2: And that’s what the article called a game-changer, right? It’s not just about making information shorter anymore. It’s about making it accessible and engaging for everyone, no matter how they learn best.

Speaker 1: Exactly. Some people are all about that audio learning, you know? And others, they might have a hard time with really dense text, so Notebook LM is trying to meet those different needs.

Speaker 2: It’s like, are we moving from lectures to this personalized conversation? It’s pretty cool to think about.

Speaker 1: Okay, so Notebook LM sounds awesome in theory, but let’s talk bigger picture for a second. The article mentioned this isn’t just some random tool, it’s part of how Google sees AI changing education. What else did you notice about that?

Speaker 2: Yeah, it’s like a whole strategy. Remember how they talked about Google for Education, like that’s already used in schools all over? And they even called out Google Classroom, which seems like it’s everywhere now.

Speaker 1: Exactly. So it seems like they’ve been kind of building towards this for a while. It makes you wonder, if they’re already in so many classrooms, could Notebook LM be the next step? Imagine if every student could use this to learn in the way that works best for them.

Speaker 2: The potential is huge. And, get this, it’s not just Notebook LM. Remember how the article talked about Illuminate, too?

Speaker 1: Oh, right, the other AI from Google. And that one’s all about making those same kind of cool conversations, but from textbooks and academic stuff.

Speaker 2: So it seems like Google’s really serious about using AI to change how we learn—not just make things easier, but actually change how we learn.

Speaker 1: It’s a totally different way of thinking about it. And, you know, it all goes back to what we were saying about how Notebook LM can create that personalized experience. Imagine a world where AI tutors are just like a normal thing.

Speaker 2: Whoa, that’s like a whole new era of education. But it wouldn’t just be for schools, would it? Like, I could use this for work or even just to learn about something new for fun?

Speaker 1: Exactly. Think about it. You could use Notebook LM to get ready for a big presentation at work or dive into a new hobby. It could really make everyone a lifelong learner.

Speaker 2: So, as we wrap things up here, I gotta ask, what do you think? Is this the future of learning?

Speaker 1: It’s definitely exciting. I really think Notebook LM is just the beginning. Like, as AI gets more advanced, who knows what kind of amazing tools we’ll have? It’s gonna be wild to see what happens.

Speaker 2: I’m with you on that one. Yeah. Well, that’s it for today’s deep dive. Thanks for exploring Notebook LM with us, and be sure to join us next time when we’ll be taking a closer look at the ethical considerations surrounding this brand-new technology.