Lectura muy interesante. OpenAI ha liberado el system card del modelo GPT-4o. Entre más potentes y versátiles son estos modelos, más problemas de seguridad acarrean, y localizarlos antes de publicar los modelos es complicado De ahí los retrasos y limitaciones de funcionalidades.
POLICY
-
Online Safety Act: Regulation Versus Censorship Debate
By
–
"Censorship" would be stopping the ability of someone to express a personal view – either online or in real life. The OSA isn't perfect. But what it's designed to do is make firms remove posts that they "disagree" with. It's about combatting harms.
-
OSA Misinformation Control Differs From Genuine Censorship
By
–
What you refer to as "censorship" is just very different to what *genuine" censorship looks like. What the OSA wants to do is limit the potential for fake news / misinformation to spread like wildfire, from one platform to another, etc.
-
UK Online Safety Act: Illegal Content and Harm Legislation
By
–
The whole point of the Online Safety Act (in the **UK**) is that it legislates against illegal content, and content intended to cause harm. You're talking about US Supreme Court ruling. We're speaking about two very different things.
-
Regulators Should Address False Claims About Real-World Events Online
By
–
I don't think it's unreasonable for ANYONE to want regulators to do something about deliberately false claims about tragic, real-world events being shared online – especially if social media companies can't achieve this themselves. Do you?
-
Anduril Secures $1.5B to Scale Autonomous Weapons Manufacturing
By
–
Anduril has secured $1.5B in Series F funding to hyperscale defense manufacturing. Introducing Arsenal-1, a 5M+ sq ft state-of-the-art facility to produce autonomous weapons systems. Learn more: https://
anduril.com/article/anduri
l-raises-usd1-5-billion-to-rebuild-the-arsenal-of-democracy/
… -
Ofcom struggles to regulate tech platforms under Online Safety Act
By
–
I'm a fan of reporting facts. Aren't you, Michael?
It's a fact that Ofcom cannot yet do its work of regulating tech platforms under the Online Safety Act.
You seem to be a fan of saying which facts should and shouldn't be reported. -
Ofcom UK Online Safety Regulations Delayed Until 2025
By
–
Under new online safety regulations in the UK, Ofcom is tasked with making sure social platforms don't allow fake news and hate speech to run amok. But Ofcom doesn't yet have the power to take action on these companies – and won't do until at least 2025.
-
Billionaire-Funded Unethical AI Propaganda Techniques Questioned
By
–
If you've got boatloads of cash from billionaires you gotta spend it on something. May as well be grossly unethical propaganda techniques eh?
-
AI-Controlled Politicians: Placeholder Humans Execute AI Decisions
By
–
9. Of course, there are some logistical issues. AIs can't run for president. But the solution is simple, and has already been tested in Japan, Denmark, Wyoming and other places: a human runs for office merely as a placeholder, and pledges to do whatever the AI says, like a