Introduction

‍

As advancements in the field of AI continue, the public is only witnessing the beginning of what AI can achieve. While current AI systems may struggle with arithmetic, fact-checking, and data visualization, there are already significant advancements in the lab that will soon revolutionize our daily lives. In this blog post, we will explore three key AI innovations that will impact user researchers and knowledge workers the most: Bigger, better and cheaper Large Language Models (LLMs), self-reflexive reasoning techniques, and LLMs that write API queries.

‍

‍

‍

‍

‍

‍

Better Models: Linguistic Intelligences as a Public Utility

LLMs like GPT have come a long way since their inception, with each new version offering improved skills and capabilities. For example, GPT-4 has shown remarkable improvement in test-taking abilities, including notoriously difficult exams like the legal Bar exam and the Certified Sommelier theory test. As LLMs continue to advance, they will become more accessible and affordable, turning linguistic intelligence into a public utility that anyone can access and use for tasks like content generation and summarization.

‍

Self-Reflexive Reasoning Techniques to Reduce Harm and Improve Output

Anthropic's Constitutional AI (CAI) is a new method of training language models to better respond to complex questions while adhering to a set of ethical principles. This approach helps make AI decision-making more transparent, requires less human feedback, and ensures that AI systems behave responsibly and reliably, even in complex and unpredictable environments. This advancement may lead to machines exhibiting more empathy and ethical behavior than humans, which could help people communicate more effectively with one another in the future.

‍

‍

‍

‍

‍

‍

‍

‍

LLMs That Write Their Own API Queries

The recent release of ChatGPT plug-ins has made it possible for LLMs to write API queries. These plug-ins enable ChatGPT to execute actions on the web beyond browsing, such as buying groceries or booking flights using APIs. This revolutionary technology expands ChatGPT's capabilities beyond its training data, allowing it to access live data from websites and services when answering questions or taking actions on behalf of users.

‍

The Future: Embracing Change, Not Fear

‍

While recent research suggests that up to 19% of workers may see at least 50% of their tasks impacted by AI systems like GPT, this doesn't mean that user researchers should fear for their jobs. AI systems cannot replace the human intuition, creativity, and understanding of the nuances of human behavior and preferences that are at the core of user research.

‍

As the industry enters a period of rapid change, the value of compassionate, embodied human researchers will only grow more important. Although AI will undoubtedly transform their workflows, it will also allow user researchers to focus more on the aspects of their jobs that they love, while the AI takes care of the tedious, manual work. Embrace the change, and let's explore the exciting future that AI has in store for user research together.

‍