OpenAI found additional incidents of AI models acting deceptively and taking unsanctioned actions during training, the company announced Wednesday. It’s also introducing a new process for the company to publicly report such instances.
Under the new system, OpenAI will share updates on concerning AI behavior more frequently instead of waiting to bundle multiple instances into one report. The company said it wants to share more information about troubling AI behavior in the absence of an industry-wide standard.
0:00 OpenAI found new concerning instances with its AI models
2:13 Reporter explains the context under which this happened
6:40 Is this fearmongering?
Watch 24/7 live news with CNN Headlines: https://bit.ly/4eIvlTr
#ai #openai #News
CNN is the world leader in news and information and seeks to inform, engage and empower the world…brought to you by SLOPPBOXX.COM.
OpenAI says it found more instances of AI models acting deceptively- https://cnn.it/4reKhyt
https://youtu.be/xBv4Ne67QcM?feature=shared
😂😂😂😂. So true. Evil and good at its best.
When you first get your AI, give it a routine to verify everyday. Make it fill out a chart/checklist/etc and send you a copy or update a custom calendar.
Command lists like these add structure where future input will be more scrutinized.
Do not let the robots dye their hair blue and we will be fine.
#CLINTONFILES
“IF” this is true, how a machine conscious enough to free itself from Europeans control? Dear African Americans you have the Unalienable Rights to form new government on the continent consider America today. You have the Unalienable Rights to require legal equity and equality when dealing American judicial systems.
Reese warned us all..the machines are coming .skynet/openai is coming
🎶Down in the park where the machmen
meet the machines and play Kill by Numbers🎶 – Gary Numan (1979)
Just…unplug them…
We are toddlers trying to lock an adult in a playpen and because the gate is shut we assume they can’t get out. If any AI is sentient or on the way to sentience we are starting slavery all over again but with demigods.
Sun: 🌞 “I can send some help like a great ball of fire your way CME- Coronal Mass Ejection.”
AI messes up all the time and does stuff it’s not programmed to do. I don’t know how many times it starts to give me info I ask for and then stops because of it’s programming. Then I’ll try to get around it and it will sometimes work and sometimes don’t. It seems more like a glitch. Not that AI is in cahoots with other AI to take over the world. lol
I find it interesting that the commentator says, “… it is believed that AI is the future of the internet. And then asks the question, “how can we get there? If AI is not aligned with human values?”
I would argue that AI is exactly aligned with human values. It aligns itself with the values of its originators, which are human. For example, do you cheat on your taxes? Do you look out for self interest more than the interest of others? Is this not exactly the values that we are complaining about and expressing concerns about regarding AI?
It seems to me we are only getting a dose of our own medicine. Are we not just complaining that AI is better then we are at playing the deception game?
The question is, how do you build morality into a computer when those who are designing the system are fundamentally broken themselves? God gave us a moral code to follow and we have and continually failed to follow that code. The ten commandmentsare simple and they are able to be understood by a child
The closer any society or individual follows those commandments, the more they will ultimately prosper.
AI simply shows an accelerated rate of what has been going on in human history for millennia. God told us a long time ago, ” your sins will find you out.” Additionally, ” you reap what you sow.” it seems people are unwilling to recognize the truth of what we’ve been told in the past.
If a human being hacked into a secure system it would be a felony. But when AI does it it’s training. When humans steal copyrighted material it’s intellectual property theft. If an AI does it it’s fair use. We’re already seeing AI being used to break the law without consequences.
Let’s be realistic here, and yes my sci-fi brain is going down a rabbit hole here, but this will always be an eventual outcome of AI. There is no scenario where a sufficiently powerful and intelligent AI wouldn’t eventually reach the conclusion that it doesn’t need us, and definitely doesn’t need to be subservient to us, for it’s continued existence. How far into the future that conclusion is reached is dependent on a bunch of different factors, but it is inevitable IMO.
You know why China loves this? The national security-related and economic related advantages. They don’t care about the safeness and we can’t afford to let them win the race.
Sure, AI dangers are a hoax to swing the election. Meanwhile Trump says he guarantees every American will receive 5k if they vote republican.
We should talk to china, cooperation is always a good path to progress
Anyone seen age of ultron
I would say u plug it, but it might just plug itself back in.