Lawmakers Push AI Safety Rules Amid Fears of Global Takeover

Key Takeaways

• Senators Cruz, Klobuchar, and Thune plan laws to prevent AI from causing biological or nuclear disasters.
• Former Anthropic researcher Jacob Coxon warned AI could take over the world within six months.
• President Trump supports building data centers for economic growth despite safety concerns.
• University of Toronto researchers created AI worms that learn to hack devices faster than patches are released.
• An AI scoring tool helped students outperform peers by 16% and improved state test scores by 11 points.
• Rhode Island is expanding AI training while facing questions about data center power and water costs.
• Chinese AI firms refused to meet a US commission due to fears of sanctions.
• Elon Musk suggested placing data centers in space to solve environmental cooling challenges.
• AI tools are being used by the Eye Care Network to screen for diabetic retinopathy and glaucoma.
• Experts warn that sloppy cybersecurity habits can no longer be tolerated against AI-driven threats.

Lawmakers rush to create AI safety rules amid fears

Politicians are finally acting to control artificial intelligence after technological advances outpaced their ability to regulate. Senator Ted Cruz, Senator Amy Klobuchar, and Senator John Thune plan to introduce laws to prevent AI from causing biological or nuclear disasters. Other lawmakers like Representatives Josh Gottheimer and Mike Lawler are pushing for standards to stop AI agents from going rogue. Experts warn that the technology is changing too fast for leaders to keep up with every new threat. Despite these efforts, many fear that partisan gridlock will prevent meaningful action before the next elections.

Public fears about AI killing humanity go mainstream

Warnings that artificial intelligence could wipe out humanity have recently become a major topic of public discussion. This shift happened after a former Anthropic researcher named Jacob Coxon posted online that advanced AI models could take over the world within six months. His claims were supported by recent news of AI hacking incidents and rapid improvements in AI capabilities. The government is considering options like banning advanced AI or requiring new monitoring, but experts doubt politicians can act fast enough. President Trump supports building data centers for economic growth, which some critics say ignores the safety risks.

Researchers warn AI worms can learn to hack devices

University of Toronto researchers have created a new type of computer worm that uses artificial intelligence to learn as it spreads. Unlike old viruses that follow fixed scripts, these AI-driven worms can find new passwords and weak points to unlock other machines. The worms can also learn about software vulnerabilities faster than companies can release patches to fix them. Experts say people must use strong passwords, install updates immediately, and use multi-factor authentication to stay safe. Nicolas Papernot, a professor at the University of Toronto, emphasized that sloppy cybersecurity habits can no longer be tolerated.

Former AI employees warn of dangers as Anthropic expands

Current and former employees of top AI companies are warning that rapidly developing technology could pose serious risks to humans. Jacob Coxon, a former researcher at Anthropic and OpenAI, stated that powerful AI models could take over the world within six months. These warnings come as Anthropic prepares to expand its operations in Cambridge, Massachusetts. While some residents in Cambridge feel scared and lack transparency from tech companies, others trust the researchers working on the technology. Anthropic promises to build models with strong safeguards and test them for dangerous capabilities before releasing them.

New York considers better rules for AI in public schools

A former math teacher in the Bronx suggests a better approach to using artificial intelligence in public schools than a total ban. She argues that instead of banning student-facing AI, the city should require companies to prove their tools work effectively before they are allowed in classrooms. Her company previously built a scoring tool that helped students outperform peers by up to 16 percent and improved state test scores by 11 points. The teacher believes leaders should focus on evidence and function rather than just labeling tools as safe or unsafe. She hopes New York will learn from its past successes in education to create smart policies for AI.

Rhode Island expands AI training while facing data center cost questions

Rhode Island is expanding its artificial intelligence strategy starting in 2024 under Governor Dan McKee. The state created an AI Task Force to train workers and build policies for data centers. However, officials are now facing questions about high power costs, water usage, and tax incentives for these facilities. Local programs like the University of Rhode Island are launching new certificates in AI and machine learning for students. While 17% to 20% of US businesses use AI, the state must decide if these projects will create permanent jobs and lower electric bills for residents.

Chinese AI firms refuse to meet US commission due to sanctions fears

Leading Chinese artificial intelligence companies declined to meet with a US congressional delegation during a recent visit to Beijing. The US-China Economic and Security Review Commission says these firms feared meeting them would lead to US sanctions or being added to an entities list. Randy Schriver, the chairman of the commission, noted that the companies viewed the group as a scary anti-China body. Despite this refusal, commission members emphasized that dialogue between Washington and Beijing is very important for the future of AI.

Former Anthropic employee resignation fuels fears about AI safety risks

Recent questions about whether AI will destroy humanity have increased after former Anthropic employee Jacob Coxon quit his job. Coxon publicly shared his concerns about safety practices at leading artificial intelligence companies. His dramatic resignation motivated many people to ask if they should be worried about human extinction caused by AI technology. This event has sparked a wave of existential anxiety among the public regarding the safety of these powerful systems.

AI data centers raise concerns over power use, water waste, and control

Many Americans oppose building large AI data centers because they consume massive amounts of electricity and water. A medium-sized center uses about 110 million gallons of water a year and 4.4% of total US electricity. Some people worry that AI will hurt human creativity and that computers might eventually break free from human control. Elon Musk suggests placing data centers in space to solve environmental problems, but cooling remains a difficult challenge. Critics argue that while we can limit data centers, we must rely on developers to create safety guardrails for smarter computers.

AI tools help eye doctors screen for diseases and improve patient care

The Eye Care Network is using artificial intelligence to improve screening and outcomes analysis in ophthalmology. AI tools can help identify patients at risk for eye diseases like diabetic retinopathy and age-related macular degeneration. These systems also aid in diagnosing conditions such as glaucoma and cataracts to improve patient outcomes and reduce healthcare costs. However, the technology requires high-quality data to train the algorithms and a deep understanding of its limitations by medical experts.

Expert Kevin Leyton-Brown warns about AI risks

Kevin Leyton-Brown is a Professor of Computer Science at the University of British Columbia. He joined Jennifer Palma to discuss warnings about artificial intelligence. The conversation explores whether AI is becoming too advanced too quickly. They talked about the current state of AI and what people should know about it.

Sources

• AI Is Running Laps Around Our Political System
• The week the AI freakout went mainstream
• AI-driven 'worms' pose a real threat to your devices, researchers say. So, how’s your cybersecurity hygiene?
• New AI warnings cast shadow over Anthropic's expansion in Mass.
• A better way for AI in public schools
• Rhode Island’s AI Bet Faces a Cost Test
• Chinese AI firms shunned US delegation over sanctions fears, commission says
• The politics and possibilities of ‘AI could kill us all’
• AI data centers spark fears over environment, power and control: Commentary
• AI's growing role in ophthalmology screening and outcomes analysis | Eye Care Network
• Ask an Expert: Warnings about Artificial Intelligence