Evaluating the Novelty of AI: A Discussion on Capability and Definition

The question of whether artificial intelligence has achieved anything "actually new" has become a focal point of debate, with critics arguing that current models merely aggregate existing information rather than generating original insights. This skepticism centers on the observation that AI outputs typically require human curation to be useful, or that they function primarily as sophisticated data synthesizers.

The Challenge of Defining "New"

Defining what constitutes a truly new contribution is difficult, as many participants in the discussion pointed out that human creativity also relies on the synthesis of past memories and experiences. Some argue that the standard applied to AI is unfairly high, as human innovation is rarely spontaneous or disconnected from prior knowledge.

"What has any human done all by itself? Everything you do comes from a synthesis of your memories of the past with your senses of the present... You think ideas appear from...what...handed to you by God through the aether from the spirit realm? We're literally meat computers."

Others suggest that if the bar for novelty is set too high, it becomes impossible to identify anything as "new" in any field, leading to a philosophical impasse regarding the nature of invention.

Scientific and Practical Breakthroughs

Despite the skepticism, some commenters pointed to specific domains where AI has demonstrably moved beyond simple aggregation. AlphaFold, for instance, is frequently cited as a breakthrough in solving the protein folding problem, representing a scientific advancement that was previously intractable for human researchers alone. Other examples include:

  • Mathematical Discovery: Projects like OpenAI's math initiatives are noted for their ability to explore formal logic and problem-solving spaces.
  • Vulnerability Research: There are reports of AI being utilized to identify software vulnerabilities, creating a new landscape for cybersecurity.
  • Efficiency Gains: Supporters argue that even if AI is not "inventing" in a vacuum, its ability to compress hours of synchronous human labor into minutes of asynchronous processing is a net-new utility in the modern workflow.

Societal and Structural Impacts

Beyond technical achievements, some observers note that AI has introduced entirely new phenomena into society, even if those phenomena are not strictly "inventions" in the traditional sense. These include:

  • The Suspicion Paradigm: The emergence of a pervasive doubt regarding the authenticity of media, where every piece of writing or imagery is now subject to scrutiny as potentially AI-generated.
  • New Problem Sets: Critics argue that rather than solving fundamental issues like climate change or housing affordability, AI has introduced new challenges, such as novel financial schemes and new forms of psychological distress.

Ultimately, the debate suggests that the frustration with AI's current state may stem from a conflation of "AI as a tool" and "AGI (Artificial General Intelligence)." As one commenter noted, the expectation for AI to act with intrinsic motivation or independent creative agency is an expectation for AGI, a milestone that has not yet been reached.

Sources