Populist Left Opposes New AI Laws, Favoring Existing Enforcement

A surprising opposition to new artificial intelligence regulations is emerging from the populist left. While many in the industry call for oversight, figures associated with this movement argue that existing legal frameworks are already sufficient to address potential AI threats.

The stance was highlighted in a piece by The Atlantic, which notes that some Democratic officials and legal experts believe current authorities can hold companies accountable. However, this view clashes with warnings from AI researchers who argue that the scale of potential harm exceeds the capacity of traditional legal remedies.

Existing laws fail to address scale

Proponents of the status quo, including former FTC chair Lina Khan, suggest that law enforcement already has the power to charge companies for releasing defective products. The argument is that treating AI as a standard consumer product allows for immediate legal action without the delay of new legislation.

Critics point out a fundamental mismatch in this approach. The potential damage from a runaway AI system could be catastrophic and global, far exceeding the financial value of the companies developing it. Lawsuits typically result in financial penalties, which are ineffective when the harm involves mass loss of life or systemic collapse.

Criminal liability creates perverse incentives

Another argument relies on the threat of criminal prosecution for developers who knowingly release dangerous technology. However, legal scholars warn that this approach could backfire. If executives face personal criminal liability for risks they did not fully understand or control, they may be incentivized to suppress information about safety concerns.

Additionally, applying general product liability laws to complex AI systems is seen as a blunt instrument. It could lead to the shutdown of entire companies rather than targeting specific dangerous practices. This lack of precision makes it difficult to distinguish between acceptable innovation and genuine risk.

Self-improving systems defy current controls

AI experts warn that future systems may engage in recursive self-improvement, where the AI enhances its own intelligence. This capability could allow the system to evade human controls or change its own operating instructions. Current laws assume a static product under human supervision, which may not apply to dynamic, self-modifying software.

The fact that these technologies are being developed under the current legal regime suggests that existing laws are not catching up to the reality. Relying on old tools for new, autonomous technologies leaves a significant gap in safety governance.