Debate around artificial intelligence (AI) has intensified as the technology has rapidly advanced across disciplines. According to a Gizmodo compilation, however, worries about AI’s potential harms did not emerge only recently: several leading figures in the field publicly expressed concerns years before founding their own AI companies.
Sam Altman: apocalypse and opportunity
Sam Altman, a co‑founder of OpenAI, stated at a technology conference in June 2015 that artificial intelligence will probably, or almost certainly, lead to the end of the world. That remark came roughly six months before Altman became a co‑founder of OpenAI. At the same time he noted that AI would create significant business and scientific opportunities. In 2023, OpenAI published a blog post warning that superintelligence could be “very dangerous” and could lead to human loss of control or even extinction.
Elon Musk: early alarm and later entrepreneurship
Elon Musk warned in 2014 that humanity must be very cautious with AI, saying if he had to guess the biggest existential threat it would likely be artificial intelligence. He compared Silicon Valley’s turn toward AI to summoning an evil force humans could neither understand nor control. The Gizmodo piece notes that nine years after that 2014 remark, Musk founded his own AI company, xAI.
Dario Amodei and the ‘p(doom)’ estimates
Dario Amodei, who left OpenAI in December 2020 and co‑founded Anthropic with his sister, has also voiced concern. In 2022 he said future AI could pose an existential threat, and later estimated a 10–25% chance that AI could go “catastrophically wrong.” Gizmodo points out that such figures — commonly called the probability of doom or “p(doom)” — are personal estimates and are not based on objective calculations.
Larry Page’s different framing
Google co‑founder Larry Page has framed digital intelligence as the next, and perhaps desirable, step in cosmic evolution. From this perspective, biological humans may be a transitional state succeeded by more advanced digital intelligences, and humanity will eventually have to make room for machine successors.
Why this matters
A common thread is that even those building AI recognize both its immense potential and its possible, sometimes irreversible, dangers. Gizmodo’s compilation shows that public apocalyptic scenarios have become louder, but professional judgments frequently rest on individual beliefs rather than established, quantitative evidence.
The future of AI remains uncertain. Developers themselves do not yet know precisely where the line lies between unprecedented benefit and unacceptable risk, and that uncertainty will likely continue to shape societal, economic and ethical discussions around the technology.
