AI News Roundup: Connecticut's AI Law, Google Gemini 3.1 Flash-Lite, and the Oscars Ban AI Actors

Connecticut Passes One of the Nation’s Most Comprehensive AI Laws

Connecticut’s legislature has approved Senate Bill 5, a sweeping AI regulation bill that passed the House 131–17 and the Senate 32–4 with strong bipartisan support. Governor Ned Lamont has said he will sign it into law, making Connecticut one of the first states with enforceable rules spanning employment AI, chatbot safety, and synthetic content provenance.

The law requires developers of AI tools used as a “substantial factor” in hiring, promotion, or termination to provide deployers with compliance information, while deployers must notify affected individuals of the technology’s use. It goes further than any other state by amending Connecticut’s anti-discrimination statutes to codify that automated decision-making is not a defense to a discrimination claim.

The bill also mandates that chatbot operators implement measures to prevent romantic or sexual interactions with minors, discourage self-harm content, and block manipulative techniques designed to foster emotional dependence. Large generative AI providers with more than one million monthly users must embed provenance data into any audio, image, or video content their systems generate. Attorney General William Tong will serve as the primary enforcement authority.

Google Makes Gemini 3.1 Flash-Lite Generally Available

Google announced general availability of Gemini 3.1 Flash-Lite on May 7, its most cost-efficient model in the Gemini 3 series. Priced at just $0.25 per million input tokens and $1.50 per million output tokens, it targets high-volume developer workloads where speed and cost matter more than maximum reasoning depth.

The model delivers 2.5× faster time-to-first-token and 45% faster output generation compared to Gemini 2.5 Flash, while maintaining similar or better quality. It scores 86.9% on GPQA Diamond, 76.8% on MMMU Pro, and has earned an Elo score of 1432 on the Arena.ai leaderboard — outperforming other models in its price tier across reasoning and multimodal understanding benchmarks.

Academy Bars AI-Generated Actors and Scripts From Oscar Eligibility

The Academy of Motion Picture Arts and Sciences has updated its rules for the 99th Academy Awards, declaring that only performances “credited in the film’s legal billing and demonstrably performed by humans with their consent” qualify for acting categories. Screenplays must be “human-authored” to compete for writing awards.

The rules don’t ban AI from filmmaking entirely — productions can still use AI tools in visual effects, sound design, and other technical areas. But the Academy has reserved the right to request detailed information about a film’s AI usage and human authorship, meaning filmmakers will need to document human involvement to maintain eligibility. The move marks Hollywood’s most significant institutional stance on AI creativity to date.

Microsoft Report: Global AI Usage Hits 17.8% of Working-Age Population

Microsoft’s latest AI Diffusion Report, published May 7, found that global AI adoption rose 1.5 percentage points in Q1 2026, reaching 17.8% of the world’s working-age population. Twenty-six economies now exceed 30% AI adoption, with the UAE leading at 70.1%. The United States climbed from 24th to 21st place with a 31.3% usage rate.

Adoption is accelerating fastest in Asia, driven by improving AI capabilities in local languages — South Korea, Thailand, and Japan saw the greatest quarter-over-quarter gains. Developer productivity continues to surge: global git pushes increased 78% year over year, and U.S. software developer employment reached approximately 2.2 million, up 8.5% from the prior year.

ServiceNow Unveils Autonomous AI Workforce at Knowledge 2026

At its annual Knowledge conference, ServiceNow expanded its Autonomous Workforce — a suite of AI agents designed to complete entire business processes without human intervention. The headline announcement was Project Arc, an enterprise desktop agent built with NVIDIA’s OpenShell runtime and governed by ServiceNow’s AI Control Tower, that will live on employee desktops and autonomously handle complex workflows.

ServiceNow also deepened its Microsoft partnership, extending AI Control Tower governance across the Microsoft Agent 365 ecosystem, and announced that all AI Control Tower capabilities are now included in every product and package by default rather than sold as an add-on. The Control Tower continuously discovers AI agents, risk-scores them, enforces least-privilege access, and measures business impact.

OpenAI Reportedly Developing AI-Agent Smartphone to Replace Apps

Analyst Ming-Chi Kuo reports that OpenAI is developing a smartphone built around AI agents rather than traditional apps, with Qualcomm and MediaTek jointly designing the custom processor and Luxshare handling manufacturing. Instead of a grid of app icons, the device would use an AI agent that completes tasks through natural conversation, combining on-device models with cloud inference.

Specifications are expected to be finalized by late 2026 or early 2027, with mass production targeting 2028. Kuo projects 300–400 million annual shipments if the device succeeds — a figure that would exceed Apple’s iPhone volumes. This is separate from OpenAI’s $6.4 billion acquisition of Jony Ive’s studio io Products, which is expected to produce AI-powered earbuds in H2 2026.