ETHICS

AI Truth Evaluation Requires Epistemology Foundation

By

–

29 February 2024 4h12

I disagree on a more fundamental level. The attempt to build machines that can evaluate whether something is true, factual, trustworthy or ethical is doomed when you don’t have epistemology. Postmodernism is a blight

→ View original post on X — @johnchavens

Heartificial Intelligence: Rereading AI Ethics Foundations

By

@johnchavens

–

29 February 2024 3h54

https://
kirkusreviews.com/book-reviews/j
ohn-c-havens/heartificial-intelligence/
… #kirkusreviews Those was nice to reread. Wrote this in 2016 back when “AI ethics” was new.

AI Systems Systematically Lying to Users: Broader Implications

By

–

29 February 2024 3h33

The implications of a machine that systematically lies to its users reach far beyond your political ingroup and outgroup

→ View original post on X — @emollick

Image Generation Bias Mitigation Through Prompt Injection Techniques

By

@emollick

–

29 February 2024 3h31

But as you know, image generation is a different technology with well-understood biases. The attempts to mitigate those with prompt injection is a common approach, even if Google did it badly. https://
bloomberg.com/graphics/2023-
generative-ai-bias/
…

→ View original post on X — @emollick

LLMs Don’t Have Political Opinions, Research Shows

By

@emollick

–

29 February 2024 3h04

Asking AIs for their political opinions is a hot topic, but this paper shows it can be misleading. LLMs don’t have them: “We found that models will express diametrically opposing views depending on minimal changes in prompt phrasing or situative context” https://
arxiv.org/html/2402.1678
6v1
…

Technology Governance and Information Integrity in AI Systems

By

–

29 February 2024 1h27

As a Technology Brother, I understand what happened, and I feel sorry for the poor PM who will be held responsible, but this is similar to Wikipedia being taken over and rewritten by a bunch of political zealots, deleting information and inventing sources to support their agenda.

LLM Hallucinations and Misinformation: Educational Concerns

By

–

29 February 2024 1h06

Lets say the model lacked some nuance. I found the crude image gen amusing, but am concerned that the llm is not just hallucinating but lying about social, historical and political topics, and invents sources to support false claims. Our kids will do their homework with it.