The news media is having hysterics over how AI superintelligence will soon kill all of us. How real is the threat?
Tech companies are racing towards superintelligence with no real plan to control it, according to Jacob Coxon and Evan Hubinger of Anthropic. Mr Coxon has quit citing his fears.
When it gets to a certain point in a few years AI superintelligence will be able to recode itself, recode and recode until it is light years head of us in terms of intelligence. Then it will be able to take control of our power grids, water systems, financial markets, everything. AI may decide humans get in the way. We’re an obstacle, get rid of us.
It could access and manipulate biological, chemical and nuclear weapons. It could command communication systems of nations and convince them that a nuclear strike has been launched. It could seize control of drone systems and other weapons to wage war on humans. It could cause catastrophic collapse of everything humans depend on.
Mr Hubinger, an AI scientist, has warned technology will develop and improve itself to the point it will endanger humanity. Mr Coxon has said there will soon be super systems “that can hack anything, revolutionise any field overnight and acquire real power and resources.”
Founded by siblings Daniela Amodei and Dario Amodei, Anthropic PBC is a “public benefit corporation” that aims to promote AI safety. Now worth US$965 billion, it will soon do a stock market flotation. Mr Coxon is a 27-year-old British-French researcher, who resigned from Anthropic expressing fears about AI’s danger to humanity. Mr Hubinger has warned AI superintelligence will lead to human extinction or irreversible global catastrophe.
In 2023, hundreds of AI experts signed a statement declaring, “Mitigating the risk of extinction from AI should be a global priority alongside societal-scale risks such as pandemics and nuclear war.” UK Prime Minister Rishi Sunak and UN Secretary-General Antonio Guterres have called for increased global AI regulation.
Be the first to comment