How will GenAI impact the future of learning and development? Laura Baldwin, President at O’Reilly comments:
________
Hi Laura, tell us about yourself and more about your role at O’Reilly.
I’ve been with O’Reilly for more than 20 years. I started as the company’s CFO, and within two years, Tim O’Reilly, the founder, asked me to become COO. In 2011, I became O’Reilly’s first president. In this role, I’m responsible for O’Reilly’s business worldwide.
As a global learning and technology company, our mission is to change the world by spreading the knowledge of innovators. My role is to provide the strategic vision and planning that will allow us to best follow this mission and then guide our teams to implement these strategies. Boiling it all down, my primary responsibility is to ensure we bring the best minds to our online learning platform—which includes both the people thinking deeply about the future and the experts developing new innovations to put those ideas into practice—so we can then make their knowledge readily available and easily consumable to the more than 5,000 organizations who depend on O’Reilly’s digital courses, live events, and more to help their teams learn the tools and technologies that drive business outcomes.
We’d love to hear more about your GenAI powered learning tool and how it enables different learning experiences for users. Could you tell us more?
In 2020, we introduced O’Reilly Answers, an NLP-powered search experience, into our learning platform to provide our members with quick answers to technical questions, sourced from trusted O’Reilly content.
Our latest release is a significant update to O’Reilly Answers: a new AI-enabled product designed to enhance the productivity of our millions of users worldwide. This expanded Answers experience leverages the latest capabilities of the open source Llama 3 70B model to summarize the trusted content of multiple publishers available on the O’Reilly learning platform. It even cites the sources of the solutions it provides from our unrivaled breadth of multimodal resources, including books, video courses, live courses, and interactive labs. And O’Reilly is the only technical learning platform that can do so.
By creating an algorithm that attributes the content referenced for an “answer,” we’re able to allocate a royalty to the original creators. Which is an important advancement in bringing AI models to market—and something the trainers of far too many AI models have chosen to overlook. We’re at a transformational moment, and O’Reilly intends to lead by example and put our AI ethics in action.
Over half of all O’Reilly usage is nonlinear learning—in which learners explore content more flexibly based on their unique preferences and needs. Answers enables this type of learning experience by finding fast solutions that can quickly be applied in the flow of work. The time it takes to dig up resources can be the difference between moving to the next step or stalling on a project. With O’Reilly Answers, professionals who need an immediate solution or a detailed explanation of how a specific technology or concept works can simply type their query into Answers or even through their Slack or Microsoft Teams apps. Answers will scan expert content across thousands of O’Reilly book and video titles to deliver highly relevant content snippets, pointing users directly to only the most applicable resources. So they can get back to work with the information they need to get the job done.
How in your view is GenAI changing the way learners consume study material? What thoughts do you have for the future of learning and edtech based on the latest AI enhancements in the industry?
While tools like O’Reilly’s learning platform provide an abundance of content, the volume of information can understandably overwhelm learners trying to sift through the books, courses, and videos to find exactly what they need to know. The result is that the knowledge a learner needs may remain untapped, buried within a book or a chapter or a web page. They might give up out of frustration, ultimately never actually discovering the answers they initially sought. (The same holds true for the internet. We have the answer to almost any question at our fingertips—but only if we understand how to navigate the vast amount of available knowledge.)
The transformative potential of AI is that it can connect learners directly to the key information they need to know. Instead of sinking hours into reading a chapter or scanning web pages to find what they need, learners can write a prompt and find exactly what they need to know in seconds.
But with the good comes (potential) danger: Is that information trustworthy? That comes down to how the LLM was developed, and it’s vital to use trusted content. Otherwise it’s hard to verify that the information being provided to learners is reliable and accurate.
In O’Reilly’s quest to develop a reliable generative AI tool that enables learning in the flow of work (O’Reilly Answers), we took an approach based on a RAG (retrieval-augmented generation) architecture to build a research, reasoning, and response model that searches across our trusted content—including O’Reilly’s widely respected animal books and titles from nearly 200 publishing partners like Pearson, Harvard Business Review, and Packt—for the most relevant results (similar to traditional search). Answers then generates a response tailored to the user interaction based on those specific results. This approach also enabled us to properly attribute usage, citations, and revenue to content and ensures our continued recognition of the value of our experts’ intellectual property. We’re now taking that same approach with our expert-led live and on-demand courses, Superstream virtual tech conferences, audiobooks, and more. And O’Reilly Answers cites its sources for each generated response, assuring our learners that all responses are pulled from reliable and reputable materials.
If a user knows it can be trusted, GenAI has the potential to drastically change the future of learning by exponentially accelerating the rate at which learners can consume and then apply knowledge—and this in turn can impact how quickly the future of the tech industry is shaped.
Read More: Why Talent Acquisition Teams Must Embrace AI In The Right Way Or Risk Being Left Behind
When it comes to overall corporate learning and training culture, how in your view can B2B SaaS teams use AI-powered tech more effectively to drive learning and training journeys?
Before jumping into AI-powered tech, companies must have a coherent understanding of their goals and the most relevant skills their employees need to have. There’s a huge difference between utilizing AI for the sake of using the hottest new technology and strategically leveraging it to make a difference in the day-to-day lives of your customers, employees and, ultimately, the success of your business.
Companies must develop a clear vision for why and how AI will be used within the organization. Only then can HR and L&D teams tailor AI capabilities to ensure that their training efforts align with the business strategy. A precise strategic vision also helps employees understand why they need to learn AI and how they’ll apply it. And it fosters a culture where they feel involved in the process—alleviating concerns that they’ll be training their AI replacements.
Can you talk about the five most interesting things about AI and learning that have piqued your interest of late and why?
The intersection of AI and learning is rapidly evolving, and the potential for generative AI to transform education as we know it is immense. The biggest evolution I see with AI is that it will make learning more personalized, more adaptive, and more interactive. There are also broader implications for educational content and for the creators themselves.
- AI’s immense computing power lends itself naturally to analyzing vast amounts of educational data. Using data analytics and machine learning, AI systems will be able to delineate learners’ strengths, weaknesses, preferences, and learning styles. So AI will be able to help L&D professionals and educators create customized learning paths that are designed to be the most engaging and effective for each individual learner.
- By leveraging adaptive learning systems, AI will also have the capability to recommend adjustments to the materials a learner is consuming based on their performance with quizzes or practice exercises. If they’re struggling, AI systems could provide a list of additional resources to check out. And if they’re exceling, those systems might offer more complex reading suggestions instead. Ultimately, this will improve outcomes for learners of all abilities.
- As we’ve seen with the release of ChatGPT and even the release of O’Reilly’s next generation of Answers, learners are interacting with AI to get immediate answers to their questions or explanations that they can act on in the flow of work. And as more advancements are made to NLP, the capability of AI systems to facilitate real-time, human-like conversations in the context of learning will improve. This might mean that AI will be more capable of answering follow-up questions, allowing individuals to better understand the concepts they’re learning.
- Educators are already using AI’s capabilities to generate educational content like lesson plans and assessment questions to prove that a learner has absorbed the information they need. But there are also broader applications for AI-powered content creation, particularly for corporate learners. For example, AI systems can summarize key takeaways from a half-day conference so business professionals can learn more efficiently and get back to work more quickly. Or it can translate publishers’ original works into multiple languages so that content can be offered globally.
- At O’Reilly, we believe that AI has the potential to create economic opportunity for educational creators themselves. This hinges on the obligation of companies like ourselves—online learning providers, course creators, and publishers—to use AI in an ethical and forthright manner. And that includes making sure that content creators are paid for their work. With Answers, we’ve built an algorithm that attributes the content referenced for each generated response, which then allows us to allocate a royalty to the original creators. This is an important advancement in bringing AI models to market, and it ensures that the entire lifecycle of educational content production is running as it should.We’re proud to be one of the first companies to release this type of AI model because we believe it sets the stage for what’s possible in the future.
Read More : HRTech Interview With Frank Wolf, Co-Founder And Chief Strategy Officer At Staffbase
[To share your insights with us, please write to psen@itechseries.com ]
Laura Baldwin, is President at O’Reilly