Google's Notebook LM: AI-powered research and communication for unions

If you've been online recently, you might have come across a recording of a couple of (fairly annoying) American voices in conversation. This pair are prolific and highly intelligent, having conversations about all manner of different topics. But recently these two remarkably versatile people had a conversations about another subject close to our heart: trade union renewal.

What’s the catch? The people are 100% AI, and their audio conversations are a feature of Google's Notebook LM, an AI research tool released by the search giant.

What is Notebook LM?

Notebook LM is Google's latest AI research assistant. It is designed to help you work with documents and research materials in new ways. You can think of it as a research partner that can read through your documents, engage in conversation about them, and help you create new content based on your materials.

But what makes it particularly interesting is its ability to transform written content into different formats - including generating natural-sounding audio discussions (aka podcasts) from text documents. The conversations created are remarkably accessible – much as human-made podcasts are for many people. They break down complex reports into formats that can be accessed and listened to on the move.

They don’t replace a full reading of the report – but do give important knowledge in a format people like and enjoy. For people with little time, having something like this available could be the difference between never engaging on something (because the report is too long and will take too long) – and having a basic overview of the topic.

How it works

At its core, Notebook LM is straightforward to use:

1. Upload documents (reports, research papers, text clips even Youtube videos, etc.)

2. Ask questions about the content or request specific outputs

3. Get AI-generated responses based on your uploaded materials

For example, we uploaded Unions21's "Roadmap to Renewal" publication from 2018 and asked Notebook LM to create an audio discussion about its key findings. Within minutes, we had a natural-sounding conversation between two voices discussing the report's main points and implications.

Opportunities for unions

Making content more accessible

One of the most exciting aspects for unions is the ability to make dense research more accessible. Consider this scenario: your union has just published a detailed 50-page report on workplace safety. Traditionally, getting members to engage with such a document can be challenging. With Notebook LM, you could:

• Generate an audio summary for members who prefer listening to reading

• Create quick-reference summaries for time-pressed reps

• Develop Q&A formats that make complex information more digestible

Enhanced research

For union researchers and policy officers, Notebook LM’s analysis tools could be very powerful, helping them:

• Quickly identify patterns and connections across multiple documents

• Generate summaries of complex policy documents

• Create briefing notes and executive summaries

• Ask questions about specific details within large documents

Challenges and considerations

While exciting, you need to consider whether using it is always the most responsible approach:

1. Data privacy: though NotebookLM won’t train on your data (so your conversations can’t end up being seen by other users), you need to consider whether the data you upload has any personal information in. Uploading personal data might contravene GDPR or other data privacy laws.

2. Accuracy: AI-generated content needs a human review. In common with all generative AI tools, Notebook LM “hallucinates” - it makes things up that were not found in your original documents. Check everything carefully!

3. Trust and transparency: What will you tell people about how you’ve used NotebookLM? Transparency about AI use is key to retaining trust. And if you’re publishing the podcast, you need to be aware that people might react badly to communications that lack human authenticity.

4. Environmental impact: Google Notebook LM is currently free to use. But that doesn’t mean there isn’t a cost involved in using it: any use of generative AI tools uses energy and water, and generates waste.

Implications for unions

Google's Notebook LM is an interesting project, as it hints at a lot of what could be really useful about AI in future for research and communications. You can collate documents for a project, and these act as a knowledge base that you can interact with. The audio conversations show us how we might create communications, on the fly, to better meet the needs of different audiences. That could be really important for unions, which have a lot of different types of messages to communicate to a lot of different types of people.

The other side of this is, of course, that for every good report that a human has slaved over – and that deserves to be more than a 50 page PDF – there’ll be someone creating thousands of different outputs for their AI-generated garbage.

While it won't replace human expertise, it offers exciting possibilities for making union content more accessible and engaging. The key will be finding the right balance - using AI to enhance rather than replace human communication, and ensuring that in our rush to innovate, we don't lose sight of the personal connection that makes union work effective.

Getting started

If you're interested in exploring Notebook LM, you should start small. Use non-sensitive documents for personal and internal use. Experiment with different output formats.

If you do start to use it more regularly, it will be important to establish guidelines for responsible use: you’ll need clear policies on what documents can be uploaded; defined review and approval processes and transparency guidelines for AI use.

If a few people are regularly using it, you might want to share learnings between teams. You can connect users and document where you’ve used the tool successfully (your use cases) to try and build expertise on where and how it can support. You’ll also want to monitor developments and updates together.

---

Note: This article is based on the beta version of Google's Notebook LM. Features and capabilities may change as the tool develops. The article was last updated in November 2024.

Image from Yasmine Boudiaf & LOTI / Better Images of AI / Data Processing / CC-BY 4.0