This link opens in a new window
While the field of AI has been around for a while, the recent developments in deep learning and neural networks have accelerated our ability to leverage vast amount of data to build better models. These models in turn, became better at things like problem solving and pattern recognition and began to set new benchmarks for natural language abilities. A recently released AI tool called Chat-GPT really turned the spotlight towards AI, as over a hundred million users flocked to test it out in the first quarter of 2023. Many are now looking to see what else this new era of tech can offer to our lives, whether that be in work or at home.
In libraries and for researchers, AI technologies hold a lot of promise to support and augment our abilities to generate, analyze, apply, and spread knowledge. Take a look at some of the videos, tools, and tips in this guide to learn more!
In the world of AI, not all tools are created equal! These apps can vary significantly in their capability and degree of specialization. You might be more familiar with those purpose-built for content generation (like Chat-GPT or Midjourney), but there are an increasing number of tools made for information retrieval and researcher workflows. How well a tool performs can vary drastically based on things like a user's input, the size and scope of its training data, and training quality (and more!). There are thousands of new apps serving niche markets and users, so experiment safely and wisely.
Here are some starting points you might wish to explore:
| Tool | Description | Notes |
| Elicit.org | Elicit can help with finding papers to cite and defining research directions. If you ask a question, Elicit will show relevant papers and summaries of key information about those papers in an easy-to-use table. It will also help you brainstorm research questions. | Sign-up required. Free and Paid versions. "How to Use Elicit Responsibly" |
| Perplexity AI | Perplexity is an AI-chat-based conversational search engine that delivers answers to questions using language models. If you ask a question, it will provide a summarized answer and can show all the sources used in-context. | Free and Paid versions. App available. |
| Semantic Scholar | Semantic Scholar is an AI-powered search and discovery tool that helps researchers discover and understand scientific literature. It will extract meaning and identify connections across papers to make it easier to find the right information. | Free. Zotero integration. Tutorials available. |
| ResearchRabbit | ResearchRabbit a literature mapping platform that takes one or more relevant seed papers and uses AI to suggest and visualize relevant literature and researchers. It can also help you stay up-to-date with new research alerts. | Sign-up required. Free. Zotero integration. |
| Consensus | Consensus is a search engine that uses AI to quickly find answers to questions from scientific research. It works by using AI to read through peer-reviewed research and extract the key findings from each paper. | Sign-up required. Free and paid versions available. |
| Connected Papers | Connected Papers allows you to input any paper identifier and create a visual graph of the related research. It uses the Semantic Scholar database (above) to source the literature. | Free (5 graphs/month) and paid versions available. |
*For a walk-through of these tools and many others, Dr. Andrew Stapleton's site and channel is a good starting-point to explore research-related tools and see how they compare!
Audit AI search tools now, before they skew research. Nature. 2023 May | Request
As we strive to improve healthcare outcomes, boost efficiency, and reduce costs, AI has the potential to play an important role in revolutionizing health care delivery. Exciting advancements are being made with medically trained large-language models like Med-PALM 2, showing us glimpses of what is to come.
However, whether we are automating administrative tasks or enhancing the accuracy of diagnostics with AI technologies it will be important to approach these emerging technologies with caution. It will be necessary to fully investigate the key risks involved and address critical privacy and security concerns, as well as establish robust ethical frameworks and explore requirements for regulatory oversight. The Canada Health Infoway AI Toolkit provides a great primer on some of these integral components as we start adopting AI technologies.
- A monthly podcast of informal conversation exploring the deep issues at the intersection of artificial intelligence and medicine
- Twice-monthly newsletter with information from the latest podcast
- Virtual events around medical AI
Next Event: "Regulating Medical AI - Staying Safe in the Fast Lane" Apr '24
