AI works through math, not consciousness

The emergence of powerful AI language models has reignited debates about machine sentience, with troubling implications for technological literacy and security. At the RSAC 2025 Conference, security expert Ira Winkler delivered a sobering reality check about AI’s true nature, emphasizing that beneath the sophisticated facades of tools like ChatGPT lies nothing more than mathematical algorithms—not consciousness or awareness—despite widespread misconceptions, particularly among younger generations.

The big picture: Three out of four Gen Z survey respondents believe AI is either already sentient or will achieve sentience soon, revealing a concerning gap in public understanding of artificial intelligence technology.

Historical context: The anthropomorphizing of computer programs dates back decades, beginning with simple programs like MIT professor Joseph Weizenbaum’s 1960s Eliza therapy bot.

• Users formed such emotional connections to the rudimentary program that they requested privacy during their “sessions,” despite the program merely reflecting their statements back as questions.
• Today’s sophisticated AI chatbots create far more convincing illusions of sentience, though they remain fundamentally mathematical constructs.

Key details: Winkler titled his conference presentation “AI is Just Math: Get Over It,” aiming to demystify AI and combat widespread misunderstandings about its nature.

• Between 90-98% of vendors selling “AI-powered” products don’t actually understand the technology they’re marketing.
• Most discussions about AI treat it as a “magical entity” rather than what it truly is: software algorithms that have existed for decades.

In plain English: AI systems—even the most advanced ones—are essentially complex mathematical formulas running on powerful computers, not conscious entities with feelings or awareness.

Important stats: Winkler identified Nvidia‘s computational hardware as a key factor enabling AI’s recent prominence.

• The technology requires massive amounts of data and processing capability to function effectively.
• Commercial algorithms like Target’s pregnancy prediction system demonstrate that “intelligence” in these systems is pattern-matching at scale, not consciousness.

Why this matters: Misunderstanding AI’s fundamental nature creates security vulnerabilities and unrealistic expectations about technology’s capabilities and limitations.

Reading between the lines: The widespread belief in AI sentience suggests a broader issue with technological literacy that could influence everything from personal privacy decisions to public policy debates.

Implications: Jobs that can be reduced to algorithmic processes face potential replacement by AI systems.

• Winkler suggests pursuing advanced mathematics degrees for greater job security in an AI-dominated future.

Where we go from here: Winkler advises specificity when discussing AI technologies rather than using generalized terminology, and maintaining healthy skepticism toward broad AI claims.