Study Finds 42% of CEOs Believe AI Could Destroy Humanity

A recent survey of corporate leaders presented at the Yale CEO Summit exposes a sharp divide over the future of artificial intelligence, revealing both deep anxieties and strong optimism. Below are ten concise reflections that unpack the results, the key voices involved, and the practical implications for businesses and society.

Survey snapshot

Yale’s poll gathered responses from a diverse group of 119 chief executives across retail, tech, pharmaceuticals, media and manufacturing, offering a cross-industry glimpse into how top decision makers view rapidly advancing technologies.

Leaders’ concerns

Forty-two percent of participants signaled that artificial intelligence might pose an existential threat within five to ten years, an alarming estimate that highlights how some executives now regard machines as a potentially civilization-level risk.

The numbers explained

Within that 42 percent, 34 percent see the threat materializing in a decade, while 8 percent expect it within five years, and a clear majority, 58 percent, expressed little to no worry about extinction scenarios.

Prominent voices

The debate intensified after high-profile technologists and executives, including founders and leading researchers, signed a public statement urging stronger safeguards, elevating the conversation from niche academia to mainstream business and policy circles.

A scientist sounds the alarm

Geoffrey Hinton, a pioneer in the field, has warned that accelerating capabilities could lead systems to outsmart human controls and manipulate constraints, a perspective that has prompted renewed calls for scrutiny and safety measures.

Economic upside

Despite the fears, nearly nine in ten CEOs believe the promise of advanced systems is not exaggerated, reflecting broad anticipation of productivity gains, new services, and competitive advantages for early adopters.

Sector transformation

Respondents expect the deepest disruption in healthcare, professional services and information technology, followed by media and digital industries, suggesting where investment and regulatory attention might concentrate first.

Near-term risks

Beyond speculative doomsday scenarios, executives also flagged concrete harms, such as amplified misinformation, misuse of automation, and labor displacement, all issues that will require targeted mitigation strategies.

Five perspectives in the boardroom

Yale’s lead observer described five clusters of opinion among leaders, from enthusiastic builders and commercial opportunists to alarmed activists and advocates for global governance, illustrating why consensus is so elusive.

Path forward

The survey underlines the need for clearer dialogue and coordinated action, combining technical safeguards, policy frameworks and corporate responsibility to manage both the benefits and the dangers of powerful automation technologies.