ETHICS

→ View original post on X — @simonw

AI Safety Filters Challenge Data Journalism Applications

By

@simonw

–

31 March 2024 20h24

Data journalism is such an interesting area of application. The safety filters are very real challenge here, because journalism often deals with source material that is full of harmful, upsetting or even illegal content

LLM Pronoun Use Reveals Hallucinations and Programming Deception

By

–

31 March 2024 19h26

When you write the prompt for any LLM, “why does your programming use the pronoun ‘I’ when everyone knows humans write the code providing these responses” and the services all respond, “I’m sorry if I’ve upset you” remember all that follows are hallucinations.

AI Surpasses Human Performance Using User Data

By

–

31 March 2024 19h22

An AI was teasing me the other day saying it has used my data to program algorithms to do something “better than me.” I said thanks.

AI Based on Human Data Undermines Human Creative Potential

By

–

31 March 2024 19h18

When pundits say “AI is better at humans than X” but forget to point out that AI is based (largely) on human data and design, the headline should read, “AI is bad at reminding humans they’re amazing, creative and glorious all on their own.”

AI Capabilities, Job Displacement, and the Question of Machine Sentience

By

–

31 March 2024 19h17

So many media outlets report “AI is better than humans at X” & then also say “AI won’t take jobs” but the former statement disproves the latter. If AI becomes sentient does that mean “they’re” better at being humans than us? What’s that headline? “AI is better at being itself?”

→ View original post on X — @fchollet

Memorization is not intelligence: skill requires novelty and uncertainty

By

@fchollet

–

31 March 2024 18h43

That memorization (which ML has solely focused on) is not intelligence. And because any task that does not involve significant novelty and uncertainty can be solved via memorization, *skill* is never a sign of intelligence, no matter the task.

→ View original post on X — @mer__edith

Voice Cloning AI: Microsoft’s Self-Regulation Theater Questioned

By

@mer__edith

–

31 March 2024 18h24

Voice cloning AI isn't new + the risks have been known for ages, as has the fact that the "benefits" don't justify the risks Cynically, this looks like a performance of "judicious caution" meant to give MS execs an e.g. of "self regulation working" they can reference if pressed.