Welcome to the PlanYear AI Newsletter! The goal: to help you understand artificial intelligence within the context of employee benefits.  Each issue, we’ll provide articles, case studies, and insights about what's going on in AI for Employee Benefits (EB).

This issue:

The Rise of “AI Native” Workflows - The biggest trend in AI we’re seeing across all industries as Q2 kicks off.

*Why this may impact EB teams - It no longer matters what format carriers use for data. As generative AI revolutionizes data intake, employee benefit teams have the potential to become independent of traditional carrier constraints. Leveraging AI to create more adaptable and personalized data structures allows teams to use the same data throughout a renewal without typing it multiple times.

How People Are Really Using GenAI - Marc Zao-Sanders, Harvard Business Review

Despite the hype around AI's potential multi-trillion dollar economic contribution and its broad user base, there's a sense of skepticism around AI. A mere 10% of knowledge workers regularly engage with ChatGPT, which dominates 60% of the market. The main reason leaders are skeptical? A perceived “lack of practical applications and inaccuracies in the technology's outputs.”

*Why this may impact EB teams - This is akin to how employee benefits teams use multiple tools in the renewal process: while the tools are helpful, their effectiveness is diminished if teams must still manually re-enter data across platforms, indicating a need for more integrated and automated solutions. This comes with a balance. According to the article, if AI is to gain trust in any industry, the ultimate goal should be to “enhance productivity and decision-making” without fully displacing human involvement, ensuring that technology serves as a “complement rather than a substitute for human expertise.”

Interesting data - Older Generations using AI:

*Why this may impact EB teams - while the prevailing sentiment among employee benefits teams we speak with suggests a generational divide in AI adoption, the statistics reveal a more nuanced picture. Despite initial assumptions that older generations might be hesitant to embrace AI technologies, the data indicates a mere 8% disparity in daily usage between younger and older age groups. This stands in stark contrast to traditional patterns of technological adoption, where newer innovations typically see a pronounced age-related skew. The nearly uniform distribution in AI usage across age demographics challenges preconceived notions about tech adoption, underscoring AI's unique appeal and accessibility. In other words, it's probably here to stay.

What’s going on in the AI world outside of Employee Benefits?

• NVIDIA has launched NIM, an agile inference microservice, enhancing the ease of deploying generative AI models across diverse environments. This tool empowers developers to experiment with an array of models, leveraging cloud-based APIs from NVIDIA's comprehensive catalog, or opt for an independent hosting solution by integrating NIM with Kubernetes for a tailored deployment. In more simpler terms (I had to google this), AI models are going to be built and deployed much faster from various environments.
• Waabi has unveiled Copilot4D, a groundbreaking foundation model tailored for autonomous driving applications. This innovative model stands out as the first of its kind, specifically designed for the physical domain, with capabilities to navigate in three-dimensional space and factor in temporal elements. Copilot4D excels in predicting how the actions of an autonomous vehicle might influence the movements and reactions of nearby traffic participants. In simpler terms - AI is now predicting how other vehicles or people might react to their actions.
• A collaboration between Stability AI and the Medical AI Research Center (MedARC), among others, has resulted in the development of MindEye2. This advanced model is capable of transforming fMRI brain scans into visual representations, requiring merely an hour of training data. MindEye2 is adept at determining the specific image a subject viewed from a set of potential images (retrieval) or reconstructing the observed image (reconstruction), complete with an accompanying textual description. In simpler terms - brain scans are now being turned into pictures and even guess or recreate the exact image a person was looking at, using just an hour's worth of brain data.