chiprook

Tech Policy and Regulation News

September 19
Policy

Former FTC Technologist Warns Against an AI 'Cartel'

Neil Chilson, head of AI policy at the Abundance Institute and former FTC chief technologist, said government-backed antitrust exemptions for AI labs could create a lasting 'cartel' that harms competition. He argued safety standards can be coordinated without suspending antitrust rules, and existing consumer protection laws already address AI harms.

Policy

Josh Hawley Wants to Haul Flock CEO's Ass in Front of the Senate Next Week

The U.S. Senate Subcommittee on Crime and Terrorism, led by Josh Hawley, will hold a September 23 hearing titled 'Always Watching: Flock's Nationwide AI Surveillance Network.' CEOs of Flock Safety, Axon, Motorola Solutions and Verkada were invited; Verkada declined to participate.

Josh Hawley Wants to Haul Flock CEO's Ass in Front of the Senate Next Week
Policy

AI Firms Knew Chatbots Were an ‘Existential Threat’ to Journalists, Court Docs Show

In a declassified court document, publishers quote Microsoft and OpenAI executives acknowledging that training LLMs on copyrighted content is the “largest theft of labor in history” and that chatbots pose an “existential threat” to the press. The document was filed in consolidated lawsuits against OpenAI and Microsoft, including suits by The New York Times and Ziff Davis.

AI Firms Knew Chatbots Were an ‘Existential Threat’ to Journalists, Court Docs Show
Policy

Elon Musk talks up AI safety while fighting regulation in wild week of strange alliances

At the All-In summit, Musk said companies should test each other’s models for safety issues before release rather than rely on regulators. According to WSJ, he, Jensen Huang and Zuckerberg advised Trump to oppose an industry AI regulator.

Elon Musk talks up AI safety while fighting regulation in wild week of strange alliances
Policy

Spanberger signs data center accountability order, creates AI task force

Virginia Governor Abigail Spanberger signed Order 22 on September 18, 2026, barring NDAs, requiring local approval for data centers above 25 MW, and ending subsidies and fast-track permitting. An AI task force will assess privacy, cybersecurity and job displacement risks.

Spanberger signs data center accountability order, creates AI task force
Policy

Why Europe has been absent from the great AI safety debate

The Guardian examines Europe's absence from AI safety debates: the continent lacks OpenAI- or Nvidia-scale giants, and slowing AI development is impossible without US and Chinese support. Lagarde and Vestager urge building domestic data centers and AI technology to cut dependence.

Why Europe has been absent from the great AI safety debate
Policy

US Federal Register removes Qwen-based search tool after FBI called Chinese models malicious

The Federal Register removed a search tool built on Alibaba's Qwen model after users noted the contradiction: the FBI calls such models a threat while a US government body used one. The open Qwen3 0.6B model ran locally and processed only public data.

US Federal Register removes Qwen-based search tool after FBI called Chinese models malicious
Policy

US and China to discuss AI governance at Trump-Xi meeting in Washington

The US and China are preparing to discuss AI governance at a Trump-Xi meeting in Washington next week. Both sides see each other as a threat: the US accuses China of technology theft, while China accuses the US of trying to monopolize AI. Experts consider a full agreement unlikely.

US and China to discuss AI governance at Trump-Xi meeting in Washington
Policy

Sony Music and UMG sue Suno, alleging new model trained with WMG and BMG infringes copyright

Sony Music and Universal Music Group filed a new lawsuit against Suno in federal court. The labels claim Suno's new model, trained in partnership with WMG and BMG, infringes copyright because it is based on a previously infringing model.

Sony Music and UMG sue Suno, alleging new model trained with WMG and BMG infringes copyright
Policy

Gavin Newsom pushes for AI kill switch in California executive order

California Governor Gavin Newsom signed an executive order on AI oversight. A panel of experts will propose recommendations within two months, including a mandatory kill switch for frontier models, independent audits, and reporting on loss of control.

Gavin Newsom pushes for AI kill switch in California executive order
September 18
Policy

Westpac becomes first NZ bank accredited to issue digital credentials

Westpac NZ became the first New Zealand bank accredited to issue digital credentials under the government DISTF system with Mattr support. Business clients can store bank credentials in the Govt.nz app wallet and share them via QR code valid for 72 hours.

Westpac becomes first NZ bank accredited to issue digital credentials
Policy

2,000 UK flight cancellations caused by previously unknown defect

A NATS investigation found that a failure in the flight data processing system on September 8 caused squawk code corruption in a millisecond, leading to a national system restart and 6-hour restrictions. 2,163 flights and about 300,000 passengers were affected, with disruptions lasting more than two days.

2,000 UK flight cancellations caused by previously unknown defect
Policy

Coders lose their DMCA case against GitHub Copilot and Codex

The US Ninth Circuit Court of Appeals upheld the dismissal of a DMCA lawsuit by programmers against GitHub, Microsoft, and OpenAI, ruling that Copilot generates code rather than reproduces it, so the demand to remove authorship information does not apply. Two breach-of-contract claims remain active.

Coders lose their DMCA case against GitHub Copilot and Codex
Policy

Connecticut bans AI health insurance denials for state workers

Starting January 1, 2027, AI cannot solely deny medical claims or cut doctor payments for ~270,000 state employee health plan participants in Connecticut; a human must confirm decisions. Insurers must disclose AI use and verify system accuracy and fairness.

Connecticut bans AI health insurance denials for state workers
Policy

Virginia’s Loudoun County Board Votes to Pause New Data Center Applications

The Loudoun County Board of Supervisors in Virginia voted 7-1 to pause accepting new data center construction applications for 12 months. The county is the world's largest data center hub with gigawatts of capacity; the pause is to assess impact on residents and revise rules.

Virginia’s Loudoun County Board Votes to Pause New Data Center Applications
Policy

Brazil's President Lula signs ReData data center bill into law

Brazilian President Lula signed bill 278/2026 establishing the ReData special tax regime for data centers: elimination of import duties, IPI, PIS/Pasep and COFINS on ICT equipment. Benefits last up to five years, then become permanent; 2026 revenue losses are estimated at 5.2 billion reais ($1 billion).

Brazil's President Lula signs ReData data center bill into law
Policy

After the Rent Cut, Your Old Solana Accounts Are Over-Funded

Solana is reducing account data storage cost by 90% in five stages; since September 11, 2026, the second step is in effect — 5080 lamports per byte instead of 6960. Accounts created at the old price hold excess funds that can be withdrawn without closing via WithdrawNonceAccount, Withdraw and WithdrawExcessLamports instructions.

After the Rent Cut, Your Old Solana Accounts Are Over-Funded
Policy

Solana’s Minimum Stake Delegation Is Now 1 SOL

On June 18, 2026, Solana mainnet activated v5 of the stake program, raising the minimum delegation from 1 lamport to 1 SOL. The program became upgradeable on-chain Core BPF bytecode, so rules change via program upgrade rather than validator release. Existing delegations below the threshold continue earning rewards but cannot be split or moved.

Solana’s Minimum Stake Delegation Is Now 1 SOL
Policy

Solana SIMD-0312 Fixes Sending SOL to Non-Existent Addresses

On Solana mainnet, the create_account_allow_prefund feature was activated on May 29, 2026, epoch 979, adding the CreateAccountAllowPrefund instruction. It allows creating an account at an address that already received SOL: the existing balance is counted, and the sender only pays the difference.

Solana SIMD-0312 Fixes Sending SOL to Non-Existent Addresses
Policy

SIMD-0500 Will Block Making Old Programs Immutable

After SIMD-0500 activation, the Solana loader will stop freezing programs with old SBPF v0–v2 bytecode: the SetAuthority instruction with authority=none will fail. As of September 17, 2026, the rule is in the validator but not activated on mainnet. To renounce program rights, it must be rebuilt with a current toolchain, upgraded, and only then frozen.

SIMD-0500 Will Block Making Old Programs Immutable
Policy

EU Kids Act Draft May Require Age Verification for Single-Player Games on Steam

The draft EU Kids Act mandates age verification for all users of online games. Under current wording, single-player games are also covered if distributed or updated via Steam. Only games on physical media without online components are exempt.

EU Kids Act Draft May Require Age Verification for Single-Player Games on Steam
Policy

Scotland Blocks AI Data Centers; Renewable Surplus Is Selling Point and Grid Trap

On September 16, the Scottish Parliament voted 80 to 26 to block all planning decisions for new hyperscale AI data centers until national planning rules are published. The reason cited is a grid crisis: curtailed renewable energy costs consumers over $1.6 billion annually in balancing payments.

Scotland Blocks AI Data Centers; Renewable Surplus Is Selling Point and Grid Trap
Policy

MiHoYo awarded $112,000 by Chinese court after AI voice service dupes Genshin Impact characters

A Shanghai court ruled illegal the use of generative AI to reproduce and sell the voices of 63 Genshin Impact characters. The defendant must stop distribution and pay MiHoYo 750,000 yuan (about $112,000); it withdrew its appeal.

MiHoYo awarded $112,000 by Chinese court after AI voice service dupes Genshin Impact characters
Policy

FDA's AI Device Rules Let Healthcare Diagnostics Skip Female Validation

A JAMA Network Open analysis of 903 FDA-approved AI medical devices found fewer than a third include performance data by sex. MIT researchers found the models recommend less intensive treatment for women than men with identical symptoms.

FDA's AI Device Rules Let Healthcare Diagnostics Skip Female Validation
Policy

FCC drops 25% foreign ownership limit for Paramount-Warner Bros deal

The FCC repealed the rule limiting foreign ownership to 25% and allowed three sovereign funds from Saudi Arabia, Qatar and Abu Dhabi to own 49.5% of the combined Paramount-Warner Bros. The commission cited the shares' lack of voting rights, but Free Press and Democrats criticized the decision.

FCC drops 25% foreign ownership limit for Paramount-Warner Bros deal
Policy

Malaysia's MyDigital ID: when optional starts to look unavoidable

Malaysia is expanding its MyDigital ID national digital identity system: since May 1 it is the only way to log into the MyJPJ app, with 12 million registrations out of a population of 35 million. App technical failures drew criticism, especially from rural residents.

Malaysia's MyDigital ID: when optional starts to look unavoidable
Policy

California Gov. Newsom issues executive order to rein in AI 'before it's too late'

Gavin Newsom signed an executive order to reduce AI risks: a panel of experts will prepare recommendations within two months on strengthening AI safety laws. Possible measures include independent audits of frontier models and a mandatory kill switch.

California Gov. Newsom issues executive order to rein in AI 'before it's too late'
Policy

EU examines Amazon over suspected price curbs on sellers

The European Union is investigating Amazon over possible unfair restrictions on companies selling on its platform, including suspicions of imposed pricing conditions on sellers.

Policy

Australia touts online harm prevention as other countries weigh it

Australian Prime Minister Anthony Albanese is traveling to the US to promote social media reform: the under-16 ban has already closed about 5 million accounts. Plans include a digital duty of care and the right to disable algorithmic recommendations, with fines above A$100 million.

Australia touts online harm prevention as other countries weigh it
Policy

AI PACs spend nearly $1 million in obscure Senate race

Super PACs tied to AI labs and investors have directed about $1 million to the Senate campaign of Republican Mike Rounds in South Dakota. Defend American Jobs, funded by an Andreessen Horowitz entity, spent about $600,000, with another $215,000 from employees and PACs linked to Anthropic, Google and Microsoft.

AI PACs spend nearly $1 million in obscure Senate race