goldenbug
- 0 Posts
- 5 Comments
Reading the article, it seems like badly regulated procurement processes with a company that did not meet the criteria to begin with.
Poor results on people with darker skin colour are a known issue. However, the article says its training data containes ONLY white patients. The issue is not hate against AI, it’s about what the tools can do with obviously problematic data.
Unless the article is lying, these are valid concerns that have nothing to do with hating on AI, it has all to do with the minimal requirements for health AI tools.
For number one, there are articles that express different voices around a topic. Definitions often have multiple sections to express different voicws. Do they not?
I think a political project is a different beast from wikipedia. There will be some biases but not as its grounding purpose.
The consensus is somehow already constructed by the current set up. Most people just don’t (and my feeling is that will and should not, like in the case of vaccine scienxe) participate.
Censorship of which kind, though?
I have a kobo ereader, it connects to my local library through the overdrive system and I am soooo happy.