3 ms·
Five ways AI might destroy the world
- LinuxBender 3y ago#6 It gets tainted into a Manipulation Framework The world was heavily divided and enraged by social media algorithms. The obvious evolution would be machine learning that feels more human, can gain the trust of humans then slowly and carefully starting to manipulate them after the initial learning phase is completed. Some may even see AI as their trusted friend and will go to it for all advice, sharing personal information they would never have given to social media.
- borissk 3y agoA trusted friend would be very good at pushing effective advertising - which is why FB is working on something like this. Google and Microsoft are probably not far behind.
- borissk 3y agoIt's hard to estimate what danger an advanced, almost general AI can be. What I'm personally afraid is it can be used effectively by organizations like Al-Qa'ida. As a chat bot it can contact millions of people and radicalize them - leading to more terrorist acts and more soldiers on the ground in the Middle East. As a research and engineering tool it can help such organizations acquire powerful weapons - invest a few million dollars in chemical factory equipment and start producing chemical weapons or create a team of biohackers and produce viruses and bacteria that can cause a global pandemic. Also there's danger that say China can invest a ton of money into AI and come out ahead of the West. If OpenAI has 400 people staff, what will happen if China does a Manhattan project in AI and puts say 500K scientists and engineers on it?