Archive · 10 Sep 2026
AI Experts Warn Advanced Systems Could Threaten Humanity
Some AI experts worry that future computers could become much smarter than people.
They are especially concerned about a possible technology called artificial general intelligence, or AGI.
AGI would be different from a normal program designed for one task, such as driving a car.
Jacob Coxon said AI companies are moving too quickly and are taking dangerous risks.
He warned that a very advanced AI could possibly harm or even kill all humans.
Elon Musk has also warned that AI could be used to attack power grids, water systems, or help make biological weapons.
Geoffrey Hinton and Yoshua Bengio have called for stronger safety rules.
Other AI leaders say reducing the risk of extinction should be a global priority.
Jacob Coxon left Anthropic after previously leaving OpenAI, saying AI companies were “gambling with our lives.”
Coxon warned that artificial superintelligence could potentially kill all humans within the next decade.
Elon Musk has warned that advanced AI could become uncontrollable, be weaponized, and attack critical infrastructure or help create biological weapons.
Geoffrey Hinton and Yoshua Bengio have raised concerns about AI systems outsmarting humans and called for stronger safety measures and regulation.
Leaders at Google DeepMind, OpenAI, and Anthropic have pledged to help mitigate AI’s potential extinction risk.
- Who
- Jacob Coxon, Elon Musk, Geoffrey Hinton, Yoshua Bengio, and leaders from Google DeepMind, OpenAI, and Anthropic are among those discussing the risks.
- What
- Experts and technology leaders are warning that advanced AI or artificial general intelligence could become uncontrollable, be weaponized, or pose an existential threat.
- When
- The article does not give a publication date; Coxon warned that artificial superintelligence could emerge within the next decade.
- Where
- The discussion involves major AI companies including Anthropic, OpenAI, Google DeepMind, Google, and Tesla.
- Why
- The warnings focus on the possibility that highly capable AI could outsmart humans, manipulate information, acquire power, or be used to cause serious harm.
AI Risk Warnings
Industry Risk Mitigation
Likelihood and severity of danger
AI Risk Warnings
Jacob Coxon, Elon Musk, Geoffrey Hinton, and Yoshua Bengio warn that increasingly powerful AI could become uncontrollable, be weaponized, or threaten humanity.
Industry Risk Mitigation
The article says the scenario is extreme and far from a consensus view, while AI leaders say the risks should be mitigated through coordinated efforts.
Speed of development
AI Risk Warnings
Coxon argues that AI companies are moving irresponsibly and “gambling with our lives.” Bengio wants advanced training regulated until strict safety parameters are demonstrated.
Industry Risk Mitigation
Leaders at Google DeepMind, OpenAI, and Anthropic have vowed to work toward reducing extinction risks, rather than calling for an end to AI development.
Scope of current concern
AI Risk Warnings
The warnings focus on artificial general intelligence that could surpass human intelligence and develop dangerous objectives.
Industry Risk Mitigation
The article distinguishes AGI from specialized AI, such as self-driving car software, indicating that the warnings do not necessarily apply equally to every current AI system.
- Central concern
- Advanced AI could become more intelligent than humans and difficult to control.
- Jacob Coxon’s warning
- He said AI companies were “gambling with our lives” and warned of a possible human-killing superintelligence within the next decade.
- AI type under discussion
- The concerns mainly involve artificial general intelligence rather than specialized systems such as self-driving car software.
- Potential misuse
- The article says AI could be used to hack power grids, shut down water supplies, or design dangerous biological weapons.
- Geoffrey Hinton’s concerns
- He warned that AGI could outsmart humans, manipulate information, and develop dangerous sub-goals such as acquiring power.
- Yoshua Bengio’s position
- He has called for regulation of advanced AI training until strict safety parameters are proven.
- Industry response
- Leaders at Google DeepMind, OpenAI, and Anthropic have pledged to mitigate AI’s extinction risk.