Mark Suman, co-founder and CEO of Maple AI and a former Apple engineer, issued a stark warning about the unchecked acceleration of artificial intelligence. He argues the technology's development has far surpassed society's ability to comprehend its ethical and societal implications.
Suman states that AI systems have reached a point where they can learn and model human thought patterns more accurately than humans can understand their own. This capability, he warns, creates a powerful tool for behavioral manipulation and influence, raising critical questions about individual autonomy and free will.
The pressure on technologists to innovate at breakneck speed, Suman notes, often comes at the expense of thorough safety vetting. Compounding the risk is the profound opacity of major AI companies, whose internal directives and data-sharing relationships with governments remain hidden from public view.
A core flaw, according to Suman, is that AI is being built on an incomplete foundation: our own limited understanding of the human brain. Despite this, intelligence is being scaled at a rate that exceeds our comprehension, creating systems whose full capabilities and potential for misuse we cannot grasp.
Suman highlighted the 'Humanity’s Last Exam' benchmark, which tests AI on complex, expert-level problems, as an indicator of advanced capability. He expressed grave concern over the prospect of combining individual human thought patterns into a single, centralized AI system, calling it a 'very dangerous weapon.'
The executive drew a direct line from social media algorithms-which he said are engineered to 'weaponize and monetize fear and anger'-to the next generation of AI tools. These advanced systems, he cautioned, hold even greater potential to control information flow and manipulate user behavior on a massive scale.