Collaborative Intelligence Week in Review - 16Jun2023

Best business article(s) I read this week…

• Understanding dark side of artificial intelligence (AI) integrated business analytics

  With everyone writing and speaking about how great AI is for businesses, the authors of this article examine the negative impacts of using AI-driven business analytics. It’s a bit academic, so you may want to skip some sections, but they make some interesting points about how specific operational characteristics limit a firm’s ability to benefit from AI.

• The economic potential of generative AI: The next productivity frontier

  On Wednesday, June 21, from 1:00 PM to 1:30PM EDT, two partners from McKinsey & Company will be sharing their thoughts on the business impact of generative AI. You can register using the link above.

Best technical article(s) I read this week…

• Macaw-LLM: Multi-Modal Language Modeling

  This is actually a GitHub repository with a ReadMe file. Describes initial attempts to synthesize leading single-purpose LLMs into a multi-modal chatbot that can interact with users about text, images, video, and audio.

• ArtGPT-4: Artistic Vision-Language Understanding with Adapter-enhanced MiniGPT-4

  Many AI researchers expect that, over the next few years, domain-specific LLMs will be more useful and valuable to organization than generalized models such as ChatGPT. In this paper, the authors describe how they modified the architecture of an existing LLM to focus on the interpretation and aesthetics of artwork.

Other item(s) of note…

• Google’s Exploration of Large Language Models in Medicine

  New England Journal of Medicine podcast featuring two research scientists from Google Health discussing their current work in adapting LLMs for clinical applications.

• On the Justified Use of AI Decision Support in Evidence-Based Medicine

  In a journal article in Cambridge Quarterly of Healthcare Ethics somewhat related to the podcast above, a researcher from the University of Copenhagen considers the question: “When is it justified to use opaque artificial intelligence (AI) output in medical decision-making?”

Coolest thing I saw…

• Fliki

  This is an AI tool for creating videos. I am exploring this for my own business. I want to produce short summary videos of longer topics. I like that I don’t have to record / edit the audio, which takes a lot of time.
  
  I haven’t spent a lot of time on it yet, but it seems pretty cool. It’s a little bit slow when processing, but I like the idea of stringing together images and brief snippets of video to make things more visually interesting.

• Microsoft Designer

  This was announced a few weeks ago, but I just got around to taking a look at it. It’s still in preview mode, but it uses AI to generate short marketing pieces. You can specify colors and fonts, so you get outputs that are more on-brand than similar tools.
  
  I created something and sent it to a Marketing Executive. She thought it was “just okay” until I told her it was AI. Then she thought it could be useful to create quick visual descriptions of pieces to give to staff or agencies for further improvements and refinements—an excellent example of Collaborative Intelligence.