As much as I think AI is incredible and I want unfettered access to it, it seems obvious that it needs to be regulated quickly. As a group of brilliant, community-oriented folks, I thought you might be interested.
NOTE
It’s important to separate content from people when discussing topics. I value open-mindedness and rational thinking and have been searching for AI future thinkers who have projected the implications into the future. After watching and listening to podcasts and YouTube videos by leading minds in AI for over 200 hours, I found this refreshing video made by Tristan Harris and Aza Raskin who are brilliant minds in their own right. Tristan was a Design Ethicist at Google for years, while Aza was behind the film The Social Dilemma. Together they founded The Center for Humane Technology. While I personally don’t care about the people behind the content as long as it makes me think, I understand that some people may feel manipulated and need certifications and guarantees to avoid being duped. If you’re one of those people, I encourage you to research these folks before forming an opinion.
This is the best content I've seen so far on AI issues:
original video at
The podcast/YouTube video discusses issues much greater than the distracting drama of the media, such as AI stealing art, AI not being considered art, AI taking jobs, and AI becoming sentient. While these are easy to understand and make for great public debate, the real dangers are far greater.
It is estimated that within six months, AI can unlock any computer lock. One way to think about it is that humans are the weakest link in security, and AI is extremely powerful at tricking humans and getting more powerful every day. For the price of a couple thousand dollars for a desktop, any half-proficient computer person can design a bioattack far worse than COVID with AI.
AI can and will be used to disseminate agendas such as political, ideological, and religious. Just think of AI getting onto platforms like eHarmony or OkCupid, with fake profile pictures, targeting vulnerable people, and convincing them of ideologies through emotions. Consider the opportunity for conning people out of money. Pre-AI, the online scam industry generated over $100 million per year.
Think about your kids' AI tutor potentially brainwashing them towards abortion or anti-abortion, guns or gun control, Christianity or atheism. Consider hackers corrupting the dataset. Anything you upload to the web right now will be consumed and integrated into the datasets, allowing you to corrupt the dataset with your agenda. I've put 120,000 photos on the web in the past three months, all of which have been consumed by data image sets for AI.
Potential risks:
Reality collapse
Fake everything
Trust collapse
Automated loopholes in law
Fake religions
Cyberweapons
Unfair lobbying
Automated fake religions
Exponential blackmail / scams
A-Z testing of everything
Synthetic relationships
AlphaPersuade
"Imagine being proficiently literate in all languages simultaneously. To take that further, this proficiency extends to not just cultural language but to mathematics, physics, chemistry, biology, python, CSS, html to in fact all languages that exist and their subsets of essoteric, exoteric and intrinsic knowledge, now then add the ability to recall how these things interrelate with one another in seconds"
I don't have any solutions, but I do want people to think about the implications and push for legislation. I dislike government interference in my affairs, but in my opinion, the consequences are too great not to push for action.
Tickled to have my chart on the growth of ChatGPT used in "The AI Dilemma" ":
Additional References
What could an AI caused existential catastrophe actually look like?
AGI and GPT Emergent behaviors - surprising
Risks from GPT-4 Byproduct of Recursively Optimizing AIs (article)
Sparks of AGI: early experiments with GPT-4 (video microsoft research)
This Just Changed My Mind About AGI ( happening fast)
interesting:
Policy research
Comments