First, congratulations to Google.org’s 20 AI for Social Good grantees, including existing Field Guide entrants Crisis Text Line, Full Fact, and HURIDOCS!

The phrase “artificial intelligence” has long been critiqued as too broad a phrase, and one that encompasses many distinct technologies. As AI becomes more relevant to our lives, we’ve made the decision to go into greater detail in our categorization of AI for good projects going forward. To do so, we’ve adapted the subcategories Google.org compiled after reviewing 2,602 AI for Impact Challenge submissions (PDF report).

We already distinguish between AI-powered bots, government AI strategies, algorithmic transparency efforts, and ethical tech projects. Now we’ll also maintain lists of projects using:

• Rules-based solutions
• Machine learning
  • Deep learning
• Audio processing
  • Speech recognition
• Machine vision
• Machine learning analytics
• Natural Language Processing

It will take some time to reclassify our AI listings into these subcategories, but our collective work to boost AI literacy in the social good sector, and society at large, will benefit from these more specific descriptions of the technologies at play.

Google.org didn’t share the full list of 2,600+ AI for social good projects, but they did summarize common themes. Here are the three “public sector” ideas they say came up repeatedly in their applicant pool:

Source: Google.org