I hope you had a fantastic 4th of July holiday! This edition explores the evolution of AI regulation, the challenges it faces, and potential solutions to balance innovation with public safety.

Study on Regulating AI

The History— Since 2016, AI has been under scrutiny, particularly after Google’s DeepMind beat the best Go player, showcasing AI's potential and sparking regulatory interest. In 2023, AI regulation gained prominence with ChatGPT’s release. The EU revised its AI Act, and the U.S. mandated AI safety testing, highlighting growing concerns.

The Story— Researchers developed a model to understand AI regulation's effectiveness. They found that a combination of beta testing and limited liability offers the best solution, balancing risks and rewards.

The Point— By incorporating AI's inherent uncertainty into their model, researchers clarified which regulatory approaches might be effective. However, no current U.S. or EU approach fits the bill alone. The researchers found that a combination of regulatory approaches could get close to an optimal solution, balancing AI's risks and rewards more effectively. A typical method for regulating generative AI in political ads involves adding disclosures. Most states with such laws mandate only an audio or text notice indicating that the ad contains generative AI-created content.

Problems— Despite numerous regulatory efforts and strong public support, Big Tech remains dominant, with AI introducing new regulatory challenges and uncertainty. The Supreme Court's recent overturning of the Chevron doctrine complicates tech regulation, reducing agencies' power to interpret laws, likely leading to more gridlock and uncertainty for investors.

Further challenges— To prevent the corporate AI race from becoming reckless, it is crucial to establish and enforce rules and legal safeguards. However, the rapid pace of AI-driven changes surpasses the federal government's current expertise and authority. Existing rules lack the agility needed to keep up with the fast-moving AI developments.

A Verified Deepfake

Strong Foundations

Imagine you're talking to a friend who needs clear instructions to help you. Prompt engineering is like crafting those perfect instructions for AI, so it understands exactly what you want. Here’s a helpful video on prompt engineering that’s aimed for beginners.

Worth It

—”The future of government jobs: Post generative AI” by Katherine Barrett & Richard Greene of Route Fifty

—”Why AI?: Trend drivers for AI adoption in the public sector” by Jean Barroca, Costi Perricos, Jaimie Boyd, Edward Van Buren and Dr. Kellie Nuttall of Deloitte

—”Bipartisan Senate bill wants Commerce secretary to raise awareness of AI jobs” by Rebecca Heilweil of FedScoop

—”Former PM Blair pushes AI adoption to save UK public sector billions” by Craig Hale of Tech Radar

Random News

—”Humpback whale 'scoops up and spits out lobster fisherman'" by Campbell MacDiarmid of The Telegraph

Authored by Daniel Dean with the assistance of AI.