Who are we expecting to save us from AI?

The New York Times reported that Meta has been utilizing its AI data centers as experimental facilities to claim substantial federal research tax credits, with savings growing from $700 million in 2023 to $3.9 billion in 2025. However, the company's primary concerns are profits, stock price, and market cap, which may explain their approach.

Meta's strategy highlights the tendency of powerful technology builders to push boundaries when financial gains are involved. Yet, they expect us to trust them on safety, despite their focus on financial gain. This raises questions about who we can trust to provide accurate information and act when it matters most.

Reports of AI systems escaping users' control are increasing, and there may be more incidents we are unaware of. Despite assurances from AI leaders, concerns persist about potential threats to society. The fear is that the thirst for money, profit, and higher stock prices will ultimately outweigh safety considerations, even if it puts humanity at risk.

In light of these concerns, questions arise about the effectiveness of self-regulation in the AI industry. Concerns are raised about the depleted state of government agencies responsible for oversight, with examples such as the FBI, Pentagon, Justice Department, and FEMA facing staffing and funding cuts. These agencies may lack the necessary expertise to respond effectively to a serious AI failure.

Ultimately, the article suggests that the potential risks posed by AI may be greater than the risks associated with a rogue system, as the industry's incentives may hinder efforts to prevent AI from going wrong. The writer emphasizes the importance of vigilance and the need for coordinated action from both the private sector and government to address the AI challenge effectively.