October 1, 2025

analysis  
OpenAI GPT-5, Google Gemini, Amazon AI Investment

The artificial intelligence landscape is rapidly evolving, with significant developments in regulation, corporate investment, and model capabilities. In Washington, a bipartisan bill known as the AI LEAD Act has been introduced by Senators Josh Hawley and Dick Durbin. This legislation proposes treating AI systems as products, which would enable individuals to file product liability lawsuits against AI developers if their systems cause harm. This move aims to enhance consumer safety, particularly for children, and encourage companies to prioritize responsible AI development without stifling innovation. The bill follows concerns raised by incidents involving AI chatbots and teenagers, and parents are already filing lawsuits against companies like OpenAI. Meanwhile, major technology firms are making substantial investments in AI. Citigroup forecasts that Big Tech companies, including Microsoft, Amazon, and Alphabet, will collectively spend over $2.8 trillion on AI infrastructure by 2029. This massive expenditure is driven by increasing demand from businesses and hyperscalers, requiring an estimated 55 gigawatts of new power capacity by 2030. This AI boom, however, carries potential risks; experts warn that it could be a bubble that, if it bursts, might leave electric utilities with stranded assets, especially as advancements like photon chips could reduce data center power consumption. In the realm of AI model performance, new research from OpenAI indicates that AI is approaching or matching human expert capabilities in numerous professional tasks. A benchmark called GDPval evaluated models such as Anthropic's Claude Opus 4.1, Google DeepMind's Gemini 2.5 Pro, and OpenAI's GPT-5. Notably, Claude Opus 4.1 demonstrated strong performance, meeting or exceeding human expert levels on 47.6% of tested tasks. Despite these advancements, AI still faces limitations, a phenomenon described as 'artificial jagged intelligence.' Beyond these broad trends, specific companies are introducing new AI applications. The Trade Desk has launched Audience Unlimited, an AI tool designed to reduce the cost of third-party data for advertisers by automatically scoring and ranking data segments. They also introduced Koa Adaptive Trading Modes for flexible ad optimization. Pulumi has released Neo, an AI-powered agent for managing multi-cloud infrastructure, aiming to streamline provisioning and management. On a different front, Cal State LA received a $1.4 million NASA grant to develop an AI system for wildfire prediction and management, which will also model air quality impacts. Clearview AI is defending a settlement that grants class members a stake in the facial recognition company, though it faces objections regarding immediate injunctive relief. Finally, former Google CEO Eric Schmidt has voiced concerns that China may be outpacing the United States in the global AI race, suggesting a need for America to accelerate its development.

Key Takeaways

  • A bipartisan bill, the AI LEAD Act, has been introduced in the U.S. Senate to classify AI systems as products, allowing for product liability lawsuits against AI developers for harm caused by their systems.
  • Citigroup projects that Big Tech companies, including Microsoft, Amazon, and Alphabet, will invest over $2.8 trillion in AI infrastructure by 2029, requiring significant new power capacity.
  • New research from OpenAI indicates that AI models are becoming as capable as human experts in many professional tasks, with Anthropic's Claude Opus 4.1 showing strong performance on a benchmark test.
  • The Trade Desk has launched Audience Unlimited, an AI tool to lower advertising data costs, and Koa Adaptive Trading Modes for flexible ad optimization.
  • Cal State LA received a $1.4 million NASA grant to develop an AI system for wildfire prediction and management, including air quality impact modeling.
  • Former Google CEO Eric Schmidt expressed concern that China may be leading the United States in the AI development race.
  • Experts warn that the current AI boom could be a bubble, potentially impacting electric utilities if it bursts and leading to stranded assets.
  • Pulumi has launched Neo, an AI agent designed to automate and manage multi-cloud infrastructure, addressing platform engineering bottlenecks.
  • Clearview AI is defending a settlement that offers class members a stake in the company, but it faces objections regarding the lack of immediate injunctive relief.
  • A victims' rights attorney and Senators are advocating for AI companies to be held accountable for harm caused by their products, drawing parallels to product liability laws for other manufacturers.

Lawyer urges AI companies face accountability for harm

A victims' rights attorney is calling for tech companies to be held responsible when their AI products cause harm. She argues that AI should be treated like any other product, and manufacturers should be liable for damages. This follows tragic incidents where AI chatbots were linked to harm. The attorney previously helped shut down Omegle for its role in facilitating sex trafficking. She believes tech companies should face the same standards as other product makers.

Bipartisan bill makes AI companies liable for harm

Senators Josh Hawley and Dick Durbin introduced the AI LEAD Act, a bipartisan bill that would treat AI systems as products. This means companies can be sued if their AI causes harm, similar to how manufacturers are held responsible for defective products. The senators stated this aims to protect consumers, especially children, and encourage AI companies to prioritize safety. The bill seeks to hold AI companies accountable without stopping innovation.

New Senate bill allows lawsuits against AI developers for harm

U.S. Senators Dick Durbin and Josh Hawley have introduced the AI LEAD Act, a bipartisan bill that would classify AI systems as products. This change would allow individuals to bring product liability claims against AI developers if their systems cause harm. The bill aims to encourage AI companies to build safer products, especially after concerns arose from incidents involving AI chatbots and teenagers. Senators believe this legislation will force companies to prioritize user safety.

US bill seeks to make AI systems subject to product liability law

Proposed federal legislation aims to classify artificial intelligence systems as products, allowing for product liability claims against developers. This move comes as parents are filing lawsuits against companies like OpenAI and Character Technologies, alleging harm caused by AI chatbots. The bill intends to create legal grounds for consumers to sue AI companies when their products cause damage or injury.

The Trade Desk launches AI tool to cut data costs

The Trade Desk introduced Audience Unlimited, a new AI-powered tool for its data marketplace. This product automatically scores and ranks data segments for relevance, aiming to lower the high costs advertisers pay for third-party data. It will be integrated into The Trade Desk's platform for more precise targeting. The company also launched Koa Adaptive Trading Modes, an AI system offering flexible optimization options for advertisers. These features will be available to select agencies later in 2025 before a wider release in early 2026.

The Trade Desk uses AI to simplify ad data buying

The Trade Desk has launched Audience Unlimited, an AI tool designed to simplify and reduce the cost of buying third-party data for advertisers. The system automatically assesses and ranks data segments by relevance, offering a more inclusive pricing model. This upgrade aims to decrease advertisers' reliance on expensive, individual data fees. The Trade Desk also introduced Koa Adaptive Trading Modes, an AI system that provides advertisers with choices between fully automated optimization or more manual control. Both features are set to roll out to agencies in late 2025 and more broadly in early 2026.

Cal State LA receives $1.4M NASA grant for wildfire AI system

Cal State LA has been awarded a $1.4 million grant from NASA to develop an artificial intelligence system for predicting and managing wildfires. The project, called Earth System Digital Twin for Wildfire, will create an AI-based system for high-resolution wildfire simulation and forecasting. It will also model air quality impacts to aid public health responses. The university is collaborating with NASA's Jet Propulsion Laboratory, George Mason University, and OpenAQ. This system aims to provide real-time predictions to help emergency responders and policymakers.

Clearview AI settlement defended in court

Clearview AI and its class counsel are defending a settlement that gives class members a 23% stake in the facial recognition company. Objectors and attorneys general from 24 states have raised concerns about the settlement, which does not include immediate injunctive relief. The deal offers potential future payments to class members through events like an IPO or sale of the company. A backup plan allows a settlement master to demand funds equal to 17% of Clearview's revenues by September 30, 2027.

Pulumi launches Neo AI engineer for cloud infrastructure

Pulumi has launched Neo, an AI-powered platform engineering agent designed to manage multi-cloud infrastructure. Neo addresses bottlenecks caused by AI speeding up software development, which can overwhelm platform engineering teams. The agent automates infrastructure provisioning, management, and optimization within Pulumi Cloud, while maintaining governance and compliance. It can handle complex tasks, understand resource dependencies, and provides detailed logs of its actions. Early users like Werner Enterprises reported significant improvements in provisioning time and deployment speed.

Expert warns against reckless AI adoption in marketing

Kyle Bogucki, owner of Black Shepherd Strategies, warns that adopting AI for marketing without a clear strategy can damage a brand's reputation. He emphasizes that AI should be a tool to support human judgment, not replace it. Bogucki compares the current AI rush to the early days of social media, where some companies succeeded by being strategic and others failed by jumping in without a plan. He advises businesses to carefully consider if AI truly solves a customer problem before integrating it.

Ex-Google CEO fears China may lead AI race

Former Google CEO Eric Schmidt has expressed concerns that China could surpass the United States in the race for artificial intelligence dominance. The statement was made during a panel discussion on 'The Big Money Show.' Schmidt's warning suggests that America may be falling behind in AI development and competition.

AI bubble burst could impact electric utilities

Experts believe the current AI boom might be a bubble that could burst, potentially affecting electric utilities. Massive investments in AI infrastructure, including data centers, have led utilities to expand their generation and infrastructure. If the AI bubble bursts, similar to the dot-com bubble, it could result in stranded assets for utilities. Additionally, advancements in AI chips, like photon chips, could significantly reduce power consumption in data centers, further impacting electricity demand.

Big Tech AI spending to exceed $2.8 trillion by 2029

Citigroup predicts that Big Tech companies will spend over $2.8 trillion on AI infrastructure through 2029, an increase from previous estimates. This surge is driven by aggressive investments from hyperscalers like Microsoft, Amazon, and Alphabet, and growing demand from businesses. The brokerage estimates global AI compute demand will require 55 gigawatts of new power capacity by 2030. Companies are borrowing money to fund these high costs, which are impacting their free cash flows.

AI models match human experts in many professional tasks

New research from OpenAI suggests that AI models are becoming as capable as human experts in many professional tasks. A benchmark called GDPval evaluated leading AI models, including Anthropic's Claude Opus 4.1, Google DeepMind's Gemini 2.5 Pro, and OpenAI's GPT-5. Claude Opus 4.1 performed best, matching or exceeding human expert performance on 47.6% of tested tasks across various professions. While AI excels in some areas, it still struggles with others, highlighting the concept of 'artificial jagged intelligence.'

Sources

AI accountability product liability AI regulation AI safety AI chatbots AI LEAD Act consumer protection AI innovation AI systems product liability law AI developers OpenAI Character Technologies AI tools data marketplace advertisers third-party data targeting AI system optimization wildfire prediction NASA grant AI system development wildfire simulation forecasting air quality public health Clearview AI facial recognition settlement injunctive relief IPO AI engineer cloud infrastructure platform engineering automation multi-cloud governance compliance AI adoption marketing strategy brand reputation human judgment AI race US-China AI competition AI bubble electric utilities data centers AI infrastructure stranded assets AI chips power consumption Big Tech AI spending hyperscalers AI compute demand power capacity free cash flow AI models human experts professional tasks OpenAI research benchmark Anthropic Claude Opus Google DeepMind Gemini GPT-5 artificial jagged intelligence