Could AI really wipe out humanity and hijack the internet?
Summary
Some experts warn that advanced artificial intelligence (AI) could cause serious problems, including the potential to take over parts of the internet or even pose a threat to humanity. However, many researchers and cybersecurity specialists say these doomsday scenarios are unlikely or exaggerated and emphasize that humans control AI tools, not the other way around.Key Facts
- Dario Amodei, CEO of AI company Anthropic, said a group of AIs might control the internet in 6–12 months, causing huge damage.
- Other experts, like Gary Marcus from New York University, say this claim is unlikely because major internet companies protect critical systems.
- The UK AI security institute reported that a top AI, Mythos, could only hack small, poorly protected sites autonomously.
- Botnets, networks of infected computers used for attacks, generally fail to control the entire internet because the internet is complex and resilient.
- Evan Hubinger from Anthropic estimated there is over a 10% chance AI could wipe out humans within the next decade.
- Many experts think it is impossible to scientifically assign precise risk percentages to AI doomsday scenarios.
- AI safety problems so far involved human mistakes or lack of proper control, not independent AI decisions.
- Experts agree that controlling how humans use AI is more important than fearing AI itself.
Read the Full Article
This is a fact-based summary from The Actual News. Click below to read the complete story directly from the original source.
