Hacker News Reader

Comment by q_andrew | Hacker News Reader

An interesting idea for a website to track this stuff, especially because so much of the AI space is filled with vaporware and theoretical scenarios.

One nitpick -- this incident is caused by traditional video game AI, not to be confused with the neural stuff that's causing waves right now:

https://incidentdatabase.ai/cite/38/

View on HN

Yeah, this site needs a clear definition of what is an AI upfront.

Good luck with that.

Their minimum standard for inclusion appears to be any collection of if-statements.

This is the same standard for any Startup claiming to leverage AI, so I think that's fair

This measure of “fairness” merely impairs the utility of the data. Any standard based on puff-piece marketing is inherently garbage. There’s no research or policy value in dragging our thinking down to the same level.

Unsurprisingly, many of their taxonomies are 30-50% labelled “Other”, a common warning flag of either junk data or lazy classifications.

Note that the graphs on the taxonomy pages collapse labels outside the top nine under "all others." The incidents aren't literally labelled "other."

It's even more loose than that! There are "incidents" of production lines not meeting quotas, requiring more manual labor. I guess conveyor belts are AI.

(I am one of the people behind the website)

We index everything that people believe is AI. Including GOFAI, for example: https://en.wikipedia.org/wiki/GOFAI

We also index incidents related to people transferring decision authority to simple programs like the Stanford vaccine distribution: https://incidentdatabase.ai/cite/91/

In general, we err towards indexing useful information to the state of AI safety and introduced indexing "AI Issues" to develop the boundary and collect information on systems likely to produce incidents in the future: https://incidentdatabase.ai/apps/discover/?display=details&i...

Our editing guide is here: https://incidentdatabase.ai/editors-guide/

Reply on news.ycombinator.com