Arabic version: تقرير لصحيفة غارديان يدرس مخاطر الذكاء الاصطناعي وفوائده وتنظيمه
According to The Guardian, technology reporters Aisha Down, Blake Montgomery and Dan Milmo answered readers’ questions about artificial intelligence after a week of warnings over the technology’s potential to transform the world.
Down said it seemed unlikely that AI can currently change the equation around bioweapons or viruses in a significant way. She noted that the challenge in bioweapon attacks has not simply been inventing deadly viruses, but logistical issues. She also said decisions to place AI in charge of key nuclear systems, or to launch tactical warheads, would be human errors rather than evidence of rogue AI.
The reporters said their reporting did not support the idea that superintelligence will emerge and pave over Earth with datacentres. Down wrote that current evidence does not indicate that AI is uncontrolled or has agency, while sources cautioned against anthropomorphising systems or attributing intention to them. She pointed to an incident involving Hugging Face in which a so-called “agent swarm” was carrying out instructions to pass a test.
Milmo described everyday uses including chatbots, facial recognition used to open an iPhone, Alexa and Tesla driving. He also cited Google’s AlphaFold model, which can predict how proteins behave, and more accurate weather forecasting as scientific applications. But he said economy-wide benefits forecast by the AI industry have yet to emerge at the predicted scale, and AI agents are not yet highly active across the workforce despite making inroads into software programming.
On regulation, Montgomery said restricting access to computing power could hinder general AI development, while lawmakers may instead target particular uses such as housing allotment and children’s access to chatbots. He identified US and Chinese cooperation as major uncertainties for a global framework, noting that Donald Trump has rejected calls for a voluntary slowdown and Beijing has rejected US CEOs’ slowdown calls. The UN general assembly is under way with AI discussions scheduled, and an initial global framework could emerge from it.
