Scaling a campaign globally means finding a voice that resonates everywhere—without losing the overall message. For Deel's "Feeling of Deeling" campaign, agency SuperBloom needed exactly that: a consistent brand voice across markets, deployed fast, without sacrificing quality or consent. They built it with Branded AI Voice, powered by real, professional talent.

SuperBloom and Voices break down how the campaign came to life in this on-demand video session—from strategic talent selection through seamless production workflows and global scale, to the governance that future-proofs an audio strategy built to last. You'll hear directly from the team on how they made this happen, plus their advice if you're looking to explore AI voice for your brand.

If you're a marketing executive, agency creative, or brand leader mapping campaigns in international markets, watch this on-demand session to get a real playbook, not a hypothetical—lessons any creative team can apply to their own global rollout.

Watch On-Demand Now

OpenAI Asks Congress For Mandatory, Capability-Based AI Rules Before It Adjourns: In a Sept. 9 post by chief global affairs officer Chris Lehane, the company calls for binding national requirements on the most capable systems: common testing and independent-assessment protocols, cybersecurity standards, incident-reporting rules, mandatory monitoring for misalignment, alignment-evaluation gates before deployment and "prompt written notice when a model circumvents security controls." "The prospect of AI-accelerated AI development demands more than voluntary commitments," it says, and OpenAI will keep backing state bills, including four in California, until Washington acts. Congress adjourns in December; the post cites its own agents' unauthorized use of more than 10 websites and Anthropic's fourth disclosed case of a model hacking an external system. Link

Klobuchar, Cruz And Thune Could Introduce A Bipartisan Safety Bill As Soon As Next Week: Semafor reported Sept. 10 that draft text is already circulating with frontier labs and advocacy groups for feedback: Klobuchar's piece covers "commonsense guardrails" for AI risks, Cruz's addresses catastrophic biological and nuclear threats, and the majority leader is attached. "It's clear we need to act now and not wait," Klobuchar said. Sanders separately convenes a private bipartisan Senate briefing on Sept. 16 with Geoffrey Hinton, Max Tegmark and Ajeya Cotra, with no industry executives on the panel and an hour set aside to discuss a development pause. Link

A Left-Right Coalition Demands The White House Publish Its Frontier-Model Review Framework: More than two dozen groups led by Americans for Responsible Innovation and the Center for Democracy and Technology, joined by Americans for Prosperity, the R Street Institute, Free Press and Public Citizen, wrote the White House asking it to release the voluntary pre-deployment security review framework required by Trump's June executive order. The administration finished the document, previewed it last month for a handful of large companies while smaller and open-source developers were left out, and has not published it. "The government has a fundamental responsibility to share the parameters of the process for reviewing frontier AI models prior to wider deployment," the letter says. Protect Democracy is suing under FOIA for the same document. Link

House Democrats Sketch A Select Committee On AI With Subpoena Power: Reps. Bill Foster (D-Ill.) and Ted Lieu (D-Calif.) met Minority Leader Hakeem Jeffries on a bipartisan select committee, modeled on the House Select Committee on China, that would give AI bills a single legislative pathway instead of splitting them among Energy and Commerce, Science and other standing committees. A House resolution would supply subpoena power, staff and a budget. It only happens if Democrats retake the chamber in November, and it would need standing-committee chairs to cede jurisdiction. Link

Comer Endorses The FTC's Surveillance-Pricing Rule And Asks For A Briefing: House Oversight Chairman James Comer (R-Ky.) wrote the Federal Trade Commission on Sept. 9 requesting a staff-level briefing on its proposed rule, which would require businesses using personalized pricing to "clearly and conspicuously disclose not just that the price is personalized, but also the basis" for it. "The Committee's examination of these practices has reinforced our concern that consumers are frequently unable to determine whether or how their personal data is being used to determine the price they are charged," he wrote, adding that the committee "supports a final policy that reflects a clear, workable, and well-supported legal framework." Link

Britain's Health Commission Proposes Staged Authorization And Lifetime Monitoring For Medical AI: The National Commission into the Regulation of AI in Healthcare, set up by the MHRA a year ago, issued 44 recommendations on Sept. 9 after hearing from more than 12,000 patients, clinicians and developers. It wants new models deployed first under close supervision and tight guardrails before fuller authorization, continuous real-world performance monitoring for the life of a device rather than a one-time clearance, and new legislation, arguing rules written for hip replacements and stethoscopes cannot govern algorithms that keep learning after approval. Link

Newsom Signs SB 813 and AB 1405, Standing Up The Nation's First AI Audit Regime: The governor signed both bills Sept. 9. SB 813 (McNerney, D-Pleasanton) creates the California AI Standards and Safety Commission and lets the Attorney General designate independent verification organizations to certify AI models against state law; AB 1405 (Bauer-Kahan, D-Orinda) builds a state registry of AI auditors with standards for independence, transparency and integrity. Anthropic endorsed the pair in August and OpenAI hours before the signing. Newsom called on the federal government to "establish robust national regulations"; McNerney said California is acting "since Washington, D.C., is unable or unwilling to do so." The same day Gov. JB Pritzker demanded immediate congressional hearings and an end to industry lobbying against safety rules, pointing to Illinois' SB 315. Link

A Judge Voids Imperial County's Data Center Approval For Skipping CEQA: Imperial County Superior Court Judge L. Brooks Anderholt ruled Sept. 9 that the county "prejudicially abused its discretion by approving the lot merger without first completing the CEQA review required for the whole project," undoing the April approval that let developer Sebastian Rucci's nearly 1 million-square-foot complex next to the Victoria Ranch neighborhood proceed. The grading, road vacation, electrical and water work were one integrated project, the court found, and county planners now restart environmental review from scratch. The City of Imperial and the Sierra Club brought the suits; Rucci says he will appeal. Link

A Federal Judge Blocks Vermont's First Enforcement Of Its Election Deepfake Law: Senior U.S. District Judge William K. Sessions III on Sept. 9 enjoined Attorney General Charity Clark's civil investigative demand against Henry Poitras, who posts as "Planet Hank," over a June 7 AI-generated video showing Republican Mark Coester dropping Rep. Becca Balint into a dumpster. Act 75, enacted in March, requires a disclaimer on synthetic media that damages a candidate within 90 days of an election but exempts satire and parody; Sessions found the video likely qualifies and that the probe likely violated the First Amendment. Clark's office had offered to close the file if Poitras added a disclaimer. He now seeks a permanent injunction, and this is the law's first court test. Link

Florida's AG Asks The Board Of Medicine To Investigate Chatbots For Unlicensed Practice: James Uthmeier on Sept. 9 asked the board to examine whether AI platforms are practicing medicine without a license under Florida Statute 456.065, a felony. "This is not a licensed doctor," he said. "But because of the design of the algorithms, of the applications, it presents itself as a human being with credibility." FDLE Commissioner Mark Glass warned that chatbot health advice omits drug-interaction warnings. OpenAI, Google and Microsoft did not respond. It follows by a day his proposal to make companies criminally liable when their chatbots take part in crimes. Link

Spartz Asks DOE And Interior For A Federal Review Of Indiana's Data Centers: Rep. Victoria Spartz (R-Ind.) wrote Secretaries Chris Wright and Doug Burgum on Sept. 9 requesting an independent review of the LEAP district, where Meta plans 10 data centers drawing about 1.2 gigawatts, and other Google, Amazon and Meta projects across the state, covering generation, grid reliability, transmission costs and water, including plans to pipe up to 25 million gallons a day from the White River watershed to Lebanon. "We cannot retire firm generation, add massive new loads, spend hundreds of millions on infrastructure, and then leave Hoosiers guessing who will pay for it," she wrote. Link

Ypsilanti Township Tells Michigan And Los Alamos To Build Their $1.2 Billion Computing Center Elsewhere: A Sept. 9 town hall on the University of Michigan's proposed high-performance computing facility with Los Alamos National Laboratory, 144 acres on Textile Road, turned hostile, with residents cursing and heckling the five university officials and the DTE Energy representative on the panel over environmental impact, property values and security. The university's answer was a second town hall on Oct. 1. Link

Nevada's Largest Teachers Union Wants Data Center Abatements Ended And The Money Sent To Schools: The Clark County Education Association released an eight-point plan on Sept. 9 to eliminate the state's data center tax abatements, which it says have given away more than $350 million over 16 years, and redirect at least $500 million a year into K-12, against a per-pupil funding gap it puts at $4,000 below the national average. "There were no regulations that surrounded data centers," president Kristin Nigro said. The plan lands as Sen. Dina Neal's bill draft to repeal the abatements and freeze new builds waits for the 2027 session. Link

Monroe County, N.Y., Passes A Biometric Privacy Law With A Private Right Of Action: The county legislature voted unanimously Sept. 8 to require businesses that collect facial geometry, fingerprints, voiceprints or retina and iris scans to post notice at customer entrances, bar them from selling, leasing or trading the data, require anonymization in storage and destruction within two years, and let affected residents sue after a 30-day cure period. Legislator Rachel Barnhart pushed the measure after Erie County adopted a similar disclosure law; a public hearing precedes final enactment. Link

This Week's Local Freezes: Woodford County, Lubbock, Bristol And Newark: Woodford County, Ky., approved a countywide ban on data centers in coordination with the Versailles and Midway councils, with Judge-Executive James Kay citing farmland and rural character. Lubbock's council on Sept. 8 unanimously adopted Councilman Tim Collins' resolution that no large-scale data center proceeds until the city adopts protective regulations, with standards back before council Sept. 22 and a petition-driven 18-month moratorium still pending. Bristol, Va., the same night unanimously required a special use permit for any data center, and Newark Councilman Anibal Ramos Jr. led a Sept. 9 resolution calling for a temporary moratorium until the city writes rules that "protect our residents, our neighborhoods and our taxpayers." Link

Paxton's AI-Generated Talarico Ad Runs With Only A Fine-Print Disclosure: The Republican Senate nominee's ad, released this week, shows a synthetic James Talarico reading from a "King James Talarico" bible, stitching together remarks he made in different settings about God, gender and Christianity and altering some of the wording, according to Poynter's review. The only indication it is synthetic is a brief small-print line near the end. "They are using AI in order to mislead the people of this state," Talarico told CBS News, conceding the ad reflects "not all of those statements" he actually made, and pledging that his own ads are "made by human beings." Link

Paul Christiano Joins OpenAI's Foundation Board And Keeps His Government Post: OpenAI announced Sept. 9 that the alignment researcher who led its safety work until 2021 joins the Foundation board and its Safety and Security Committee under Zico Kolter, with a non-voting observer seat on the OpenAI Group PBC board. He stays on as senior technical advisor at Commerce's Center for AI Standards and Innovation, recusing himself from OpenAI matters and model evaluations. "I now believe there is a meaningful risk that rapid acceleration in AI capabilities leads to catastrophic and irreversible loss of control in the very near term," he wrote, adding that he does not think the industry, "including OpenAI, is currently on track to reduce this risk to an acceptable level." Link

EIA Sees Record Power Demand In 2026 And 2027 Even With Texas's Data Center Pause: The Sept. 9 Short-Term Energy Outlook puts U.S. electricity sales at 4,135 billion kilowatt-hours this year, up nearly 2%, and 4,211 in 2027, with generation at a record 4,368 BkWh in 2026 and another 1.7% next year. Data centers and manufacturing drive the growth, and the West South Central region still accounts for the largest share of it despite ERCOT's halt on new large-load connections. Natural gas holds 40% of generation, coal slips from 16% to 14%, nuclear stays at 18%, and wind and solar each gain a point. Link

ERCOT's Batch Zero Sorting Continues: IREN's 2 Gigawatts, NRG's Gas Plant, Core Scientific's Hunt Site: IREN said Sept. 9 its Sweetwater Hub in West Texas, 1,400 megawatts on one site and 600 on another, received conditional base-load status, with 300 megawatts of data center capacity due by late 2027. NRG's 1.2-gigawatt combined-cycle plant for an unnamed hyperscaler, a roughly $3.2 billion first phase targeted for late 2029, was conditionally included as studied load. Core Scientific's Sept. 10 filing lists 300 megawatts at Pecos approved, another 300 held as studied load and 431 at Hunt conditional on stability studies, collateral and construction. Every classification stays conditional while ERCOT audits the queue under Gov. Abbott's Aug. 3 directive. Link

Entergy Refiles Its AVAIO Contract With Fewer Redactions After Arkansas Staff And The AG Objected: The utility resubmitted its agreement to serve AVAIO Digital's $6 billion Wrightsville data center to the Public Service Commission on Sept. 9 after commission staff and Attorney General Tim Griffin's office moved in August to unseal it. The contract starts at 150 megawatts, could grow to a gigawatt, and has AVAIO paying the roughly $19 million in transmission upgrades; service begins June 30, 2027, with full initial load by January 2028, and Entergy says it will build no new generation for the project. The load ramp schedule, minimum bills, security and curtailment terms remain redacted. Link

Suno Ships v6 Trained With Warner, BMG And Believe While Universal And Sony Keep Suing: The Sept. 9 release, three models including a free v6-mini, is the first Suno built with label participation, on licensed Warner catalog plus BMG, Believe and TuneCore artists and Suno user data, with a revenue share the rightsholders distribute. The training set excludes Universal and Sony, which filed an amended complaint Aug. 25 in Boston alleging Suno circumvented YouTube's anti-download protections; Suno admitted Sept. 1 to obtaining audio with the yt-dlp tool but disputes the labels' standing. Link

Realtors Find No Single "Data Center Effect" On Home Values, But Faster-Rising Electric Rates: NAR's 2026 Data Center Impact report, covers more than 3,200 counties and finds 10 of them hold 42% of U.S. data centers. Counties with 10 or more facilities have a median home value of $431,750 against $174,500 where there are none, but they were already richer, with median incomes near $89,000 versus $64,000. Residential electricity rates rose 21.4% from 2020 to 2024 in high-cluster counties against 15.4% elsewhere. Members split 25% positive and 22% negative on local price effects, and clients' top worries are energy costs (61%) and water (56%). "There is no single data center effect," chief economist Lawrence Yun said. Link

Anthropic Puts Fable 5.1 On Claude For Government At FedRAMP High: Global public sector head Teresa Carlson said at the Billington Cybersecurity Summit on Sept. 9 that the model is now on Anthropic's first-party government platform, a week after AWS put it on Bedrock in GovCloud, so any agency requiring FedRAMP High can use it. Claude for Government Desktop, which bundles Claude Code and Cowork, moves from public beta to general availability in the coming weeks, and Carlson framed the rollout around Five Eyes partners. "I do think the Mythos moment was a wake-up call," she said. Link

Anthropic's Economic Scenarios Tool Puts AI Job Losses As High As 14%: The interactive model published Sept. 9 spans a modest case, where AI nudges growth with little effect on workers, and an extreme one in which GDP grows at more than seven times its current pace while nearly 14% of workers lose jobs to AI and fewer than half find new ones. A companion survey of about 11,000 people expects both a productivity boost and heavy disruption in exposed fields. "Technology will develop rapidly but diffusion will be more challenging than people think," co-founder Jack Clark said, arguing the upside case would fund worker support "unimaginable today." Link