SAIL: State of the AI World, Ethics, Tools
The goal of this newsletter is to track trends related to AI that have an impact on learning and knowledge processes.
This is no easy task. The world of AI moves at a dizzying pace. Understanding the small scale innovations, such as new advancements in hardware or new algorithms that outperform previous generations by a small percentage, are the domain of experts. Broad general trends can be captured by annual reports such as this one by Stanford's HAI. The trends detail the state of AI in industry, skillsets, and adoption. I had a brief moment of joy when I say chapter 5 (AI and Education), but it ended up being about the state of education in AI. While that is obviously an important topic, I'm more interested in what we can expect AI to do to the education sector. Still, great overview report!
Equity questions are front of mind in all areas of society. It's commonly understood that data are not neutral and bias exists in analysis (from data collection through to cleaning through to running algorithms). Another important challenge exists, however, economically: not all systems have the ability to be involved in cutting edge research. A recent report addresses exactly this concern: "unequal distribution of compute power in academia is furthering inequality in the era of deep learning." Unfortunately, it doesn't stop there. Universities are losing top researchers to industry.
Google, however, is eager to help. Ethics as a service. Not sure why, but something is unsettling about that.
Here's a fun site: AI Tools. Go play.