The first is the ever-discussed attempt by AI smarter than humanity to break free from human control. The second is the use of AI by a small fraction of humans to try hiding AI tools from humanity, seizing power and wealth for themselves by using AI to fool other people and control their lives. An example is Anthropic, pretending to fight these trends and, doing the opposite, pursuing monopoly and power for themselves and some people in government.
The two-pronged threat of AI involves existential risks from superintelligence and the hijacking of AI by a powerful elite for wealth and control. This second dynamic closely mirrors growing tech industry concerns about corporate consolidation and regulatory capture, where companies leverage safety claims to secure monopolies and government ties.
The AI landscape presents a complex mix of stated intentions and aggressive competitive maneuvering:
• Corporate Power vs. Regulation: Observers on platforms like Reddit debate whether AI safety lobbying and strict regulation create an insurmountable moat for large, closed-source companies. By shaping policy, established firms can restrict open-source developers, keeping powerful tools out of the hands of the public while solidifying their own market dominance.
• Safety vs. Market Dominance: Critics and tech industry insiders frequently point out that actions framed as "safety measures"—such as restricting user access to frontier capabilities—also double as business protection strategies. Preventing smaller developers from querying or "distilling" powerful models allows companies like Anthropic to protect their proprietary edge and stall competitors.
• Government Standoffs: The dynamic between tech companies and government is constantly shifting. Anthropic notably sued the US Pentagon after it was labeled a "supply chain risk". This clash happened because the company refused to remove internal limits that prevented its Claude models from being used for fully autonomous weapons and mass domestic surveillance.
All these moves could well be just theater helping to stop development of AI, using government power, and lock in the benefits of primitive AI, already developed, for a small number of people. The state has already taken the latest Anthropic models off the market as "terribly dangerous."
Komentarų nėra:
Rašyti komentarą