What to Watch in AI: The Future of Search Engines

What to Watch in AI: The Future of Search Engines

In today's fast-paced and knowledge-driven world, finding the information we need quickly and efficiently is crucial. However, the traditional way of searching for content at work is becoming increasingly broken. With the exponential rise in knowledge and the distributed nature of work, it is becoming more challenging to locate existing knowledge. This is where intuitive work assistants like Glean come into play, transforming from a nice-to-have tool to a critical component in driving employee productivity.

As organizations become more distributed and knowledge becomes fragmented, the need for appropriate governance controls becomes paramount. Enterprises often struggle with enforcing these controls, such as understanding what the end user is allowed to see, ensuring the inference is done on their servers, and identifying the source data that led to a given model output and who owns it. Overcoming these obstacles is essential for enterprises to ship AI applications to production successfully.

One of the most tedious and expensive parts of the AI process is data processing and annotation. However, it is also the most crucial for achieving high-quality outcomes. Despite the availability of pre-trained large language models, enterprises must prioritize using their proprietary data across multiple modalities to create production AI that offers differentiated services, valuable insights, and increased operational efficiencies.

The future of search engines lies in Generative AI. The current search engine design, both on mobile and desktop, is based on technology from the late 1990s. While it has served us well thus far, the content we consume has drastically evolved. From social networks and data streams to video content and authoritative knowledge, our consumption habits have changed significantly. Instead of relying on a static database and searching for information, the future of search lies in leveraging that database as training data and generating results with neural networks.

Trained models are incredibly small compared to the vast amount of training data available. For example, Stable Diffusion is only around 2 gigabytes, whereas the training data amounts to 100 terabytes. The shift towards generative AI search means that users will no longer have to search for something and sift through multiple results. Instead, they will be able to generate the answer they are looking for directly. This new approach will completely transform the user flow and eliminate the need to navigate through countless pop-ups, ads, and scams.

While this paradigm shift in search engines holds great promise, it also raises questions about the future of the advertising business. In a generative AI search world, the traditional distribution monopoly and advertising model of incumbents will be bypassed. Training a model may be expensive initially, but the cost of running it is negligible. This could potentially disrupt the current advertising landscape and force advertisers to explore new avenues for reaching their target audience.

In conclusion, the future of AI and search engines is evolving rapidly. To stay ahead of the curve, organizations must adopt intuitive work assistants like Glean, enforce appropriate governance controls, and prioritize the use of proprietary data for AI development. Additionally, the shift towards generative AI search presents both opportunities and challenges for the advertising industry. As AI continues to advance, it is crucial for businesses to embrace these changes and adapt their strategies accordingly.

Three actionable pieces of advice for organizations looking to leverage AI for search and knowledge management are:

1. Invest in intuitive work assistants: Incorporate tools like Glean into your organization to optimize employee productivity by providing a seamless and intuitive search experience.

2. Prioritize data governance: Establish robust governance controls to ensure compliance, privacy, and ownership of data used in AI applications. This will not only protect your organization but also build trust with users.

3. Embrace generative AI search: Explore the potential of generative AI search engines to revolutionize how users access information. Stay ahead of the curve by experimenting with neural networks and training models to generate accurate and personalized search results.

As we venture into the future of AI-driven search engines, it is essential to adapt and embrace the changes that come along. By leveraging the power of AI, organizations can unlock new possibilities, enhance productivity, and create innovative solutions that cater to the evolving needs of users.