Breaking AI & Tech News Analyzed

The Lawsuit Against Meta California’s Santa Clara county has sued Meta Platforms, alleging it has profited from Facebook and Instagram ads promoting scams in violation of California’s false advertising and unfair business practices laws. Allegations of Tolerating Fraudulent Advertising The lawsuit – filed on Monday in Santa Clara county superior court on behalf of all California residents – accuses the social media giant of tolerating fraudulent advertising on a global basis. The suit seeks restitution, civil damages and an order prohibiting Meta from engaging in unfair business practices. Revenue from High-Risk Scam Ads Citing leaked internal documents first reported by Reuters last year, the complaint alleges that the company earned as much as $7bn in annual revenue from so-called “high-risk” scam ads which show clear signs of being fraudulent. Meta's Response and Defense Meta said it intends to defend itself against the claim. “This claim relies on Reuters reporting that distorts our motives and ignores the full range of actions we take to combat scams every day,” said a Meta spokesperson, Andy Stone. “We aggressively fight scams on and off our platforms because they’re not good for us or the people and businesses that rely on our services.“ The Impact on Users and the Legal Proceedings In the suit, Santa Clara alleges that Meta materially contributed to an epidemic of fraud by allowing middlemen to sell accounts to place ads that were protected against enforcement, and targeting scam ads at users who had clicked on similarly bogus offerings in the past. The county will retain full control over decisions involving the case, and outside law firms will only be paid if the county wins.

Introducing Poppy's Proactive AI Assistant Poppy, a new app, aims to organize the chaos of digital life by combining calendar, email, messages, and other sources into a single dashboard. The app uses AI to guess what's important to users based on their data and provides proactive suggestions to help them manage their time. How Poppy Works Users can connect various services to Poppy's app, such as email, calendar, and location. Poppy then uses this data to provide suggestions, such as taking a break to go for a walk or suggesting restaurants based on a friend's food preferences. Key Features and Integrations Works with everyday apps like Apple Calendar, Google Calendar, Gmail, Outlook, and more Integrates with apps like Uber and Instacart Users can message Poppy with questions or requests Poppy can track flights and alert users to changes Security and Future Plans Poppy's maker, Sai Kambampati, emphasizes data security, with encryption and a zero-retention policy. The company plans to switch to local, on-device AI models in the future. Funding and Team Poppy's San Francisco-based team of four is backed by $1.25 million in pre-seed funding led by Kindred Ventures, with various angels participating, including DeepMind's Logan Kilpatrick.

The Regulatory Response to Deadly Online ContentA nihilistic internet suicide forum implicated in over 160 UK deaths has been fined £950,000 by the online regulator in its latest attempt to shut it down. Ofcom said the US-based website remains accessible in the UK despite over a year of warnings. Online safety campaigners have accused the regulator of taking an "interminable" amount of time to act.The Samaritans, mental health campaigners and the Molly Rose Foundation, have repeatedly raised concerns about the site – which promoted a particular poison – which has remained accessible despite it being cited in multiple coroners' reports regarding the deaths of UK citizens.Molly Rose was set up in the memory of Molly Russell, a 14-year-old who took her own life after descending into a vortex of negative online content, including about suicide. Its chief executive, Andy Burrows, welcomed the fine and a separate move that could block UK internet access to the site, but said it was "appalling that it has been left to bereaved families and campaign groups to press Ofcom into action".Legal Action Under the Online Safety ActOfcom has been trying to get the site to obey British laws criminalising intentionally encouraging or assisting suicide since last spring. It had some success with the site being blocked last July and then a mirror site being taken down in November. But it is now taking action because the site can be "used by people in the UK, including without a VPN, and presents a material risk of significant harm".The fine is being levied under the Online Safety Act which also allows Ofcom to seek a court for an order requiring internet service providers to block UK access to the site. It is preparing an application to have its connections effectively cut "if our concerns are not fully addressed and there continues to be an ongoing breach".It accused the provider of "serious and deliberate contraventions" and said the fine reflected the "the risk of fatal harm to people in the UK posed by the content present on the service".Human Cost and Family AdvocacyOn Wednesday, the forum was unavailable but its operator posted a page that said it was advocating for "the right to access lawful information without government overreach" and quoted Mark Twain: "Censorship is telling a man he can't have a steak just because a baby can't chew it."The Molly Rose Foundation and Together with Families and Survivors to Prevent Online Suicide Harms said coroners had warned the UK government 65 times about risks of further deaths from the forum "and a substance it promotes, glorifies and instructs for use as a suicide method". Adele Zeynap Walton, the sister of Aimee Walton who took her life after accessing suicide forums, said the wait for action had been agonising."While we've waited further lives have been lost and we've had to fight every step," she said, speaking on behalf of Families and Survivors to Prevent Online Suicide Harms. "We feel let down by the process and Ofcom's slow response to this threat to life."Regulatory Challenges and Future ActionsBurrows said: "Molly Rose Foundation submitted detailed evidence which showed scores of vulnerable young people remained at risk while Ofcom's investigation dragged on … There are real questions about why it has taken so long for the regulator to act against a forum linked to at least 164 UK deaths."Ofcom said it has "engaged extensively" with the forum provider and that last summer the forum geoblocked mirror sites in the UK and later removed information on a landing page promoting ways to circumvent the block."We share the urgency about the extreme harms that sites such as this can cause, and understand the anger felt towards them by those who have been so personally affected," a spokesperson said. "It is vital that we ensure our enforcement action is thorough, and this can take time, as is the case for any enforcement agency."Preventive Measures and Support Resources"Lucas was 16. Vlad 17. Aimee 21. Grace, Hannah and Tom 22. Immy 25. Adam 28 and Claire 41," the campaign groups said in a report last year. "They were drawn into a dark world that was allowed to exist online and continues to exist through the use of a VPN. We believe our loved ones suffered coercion, grooming, instruction on how to end their lives. Most accessed a poison that was allowed to cross borders or was readily available domestically."Suzanne Cater, director of enforcement at Ofcom, said the forum had "caused unimaginable pain and suffering … and no punishment can undo that harm"."The provider of this forum knows it's used to share illegal content encouraging and assisting suicide on their site," she said. "While they've responded to our enforcement action by making some changes to the accessibility of their service in the UK, this is not good enough and the changes they've made were not consistently applied or effective to reduce the risk of harm. Given the ongoing risk of harm, we are using all powers available to us to protect the public."

Adaption announced the launch of AutoScientist, a new AI‑training product that automates fine‑tuning and data optimization to help frontier models acquire specific capabilities faster. Co‑founder and CEO Sara Hooker highlighted the system’s ability to co‑optimize both data and model, positioning it as a potential catalyst for broader, outside‑lab AI breakthroughs. AutoScientist: Automated Fine‑Tuning for Faster Capability Gains The platform builds on Adaption’s existing Adaptive Data service, turning continuously improving datasets into continuously improving models. By automating the conventional fine‑tuning workflow, AutoScientist aims to make high‑quality model adaptation a plug‑and‑play process for a wide range of domains. Performance Claims: Doubling Win‑Rates Across Models Launch materials state that AutoScientist has more than doubled win‑rates across different model families. Traditional benchmarks such as SWE‑Bench or ARC‑AGI are not directly applicable due to the tool’s task‑specific adaptation focus. The service is free for the first 30 days to encourage early adoption and real‑world validation. Strategic Implications for Frontier AI Labs By reducing the manual effort required for data curation and model fine‑tuning, AutoScientist could lower the barrier to entry for labs aiming to train cutting‑edge models. This aligns with the broader industry trend of “neolabs” leveraging heavy investment to accelerate self‑improving AI research outside of traditional corporate labs. Future Outlook: Open Access and the Race to Self‑Improving Models If the promised performance gains hold up in practice, AutoScientist may become a standard component of AI development pipelines, spurring faster iteration cycles and potentially democratizing access to frontier AI capabilities. Hooker predicts that, similar to the impact of code‑generation tools, this platform could unlock a wave of innovation across multiple fields.

New research from the International Data Center Association (IDCA) reveals that datacentres are now responsible for roughly 6% of electricity consumption in the United Kingdom and the United States, intensifying public and political scrutiny over the sector’s rapid energy growth.Datacentre Power Demand Hits 6% of UK and US GridsThe study notes a 15% worldwide increase in datacentre electricity use over the past two years, driven by the surge in AI workloads and internet traffic. Annual global investment in new facilities is approaching $1tn (£740bn), equivalent to nearly 1% of the global economy. In the UK, datacentre electricity share has risen to 5.9%, while the US sits at 6%, far above the global average of 2%. Smaller nations such as Singapore and Lithuania face even higher pressures, with datacentres consuming 19% and 11% of their national grids respectively.Financial and Energy Metrics Highlight Rapid GrowthGlobal investment: ~$1tn in 2025UK grid‑connection queue: grew 460% in H1 2025US “zombie” services: account for 13% of datacentre load, equating to over 3 GW of wasted powerProjected UK demand: could quadruple by 2030These figures align with the International Energy Agency’s estimate that global energy use by datacentres rose 17% in 2025, outpacing overall electricity demand growth of 3%.Community Pushback and Policy Implications Across NationsThe IDCA warns that once a country’s datacentre footprint reaches the 5%‑6% threshold, “significant community and political pushback” becomes inevitable. In the UK, activists and groups such as Greenpeace UK have warned of an “unchecked AI boom” leading to higher energy bills, water‑stress, and renewed reliance on fossil fuels. The report calls for:Greater transparency from tech firms on future datacentre plansMandatory environmental impact assessmentsA ban on new polluting power plants dedicated to AI workloadsAdditionally, the study highlights emerging security concerns, noting that recent attacks on datacentres in the Middle East have underscored the need for integrated cyber‑physical protection strategies.Outlook: Regulation, Transparency, and Security Challenges AheadLooking forward, the IDCA predicts that pressure will mount for:Stricter national grid connection policies to curb the 460% surge in pending requestsIndustry‑wide standards to eliminate “zombie” services and improve energy efficiencyCoordinated security frameworks that address both cyber threats and physical vulnerabilitiesIf policymakers act swiftly, the sector could mitigate its environmental footprint while sustaining the growth of AI and cloud services. Failure to do so may trigger broader societal resistance and accelerate regulatory clampdowns.

The humanitarian crisis in Haiti's capital has escalated dramatically, with residents of the Cite Soleil neighborhood taking to the streets to demand immediate government protection after a weekend surge in gang violence forced hundreds to flee their homes. The Collapse of Security in Cite Soleil Protesters reported witnessing people being killed in Cite Soleil in recent days, yet Haitian authorities have remained silent regarding casualty figures. This violence is part of a broader trend where armed gangs have tightened their grip on the capital since the assassination of President Jovenel Moise in July 2021. Police reports indicate these groups now control approximately 70 percent of the capital. Displacement Statistics and Gang Dominance The displacement crisis has reached critical levels. A report by the International Organization for Migration estimates that more than 1.4 million people have been uprooted by gang warfare. Of these, about 200,000 are currently living in overcrowded and underfunded sites within the capital, highlighting the severe strain on urban infrastructure. The Humanitarian Toll on Healthcare Medical services are facing an existential threat. In a statement released on Monday, Doctors Without Borders (Medecins Sans Frontieres) announced the evacuation of its hospital in Cite Soleil following intense clashes on Sunday. Furthermore, the Centre Hospitalier de Fontaine suspended operations, having successfully evacuated all patients, including 11 newborns, to safety. The Unfolding International Security Mission The situation is poised for a potential shift with the deployment of an international security mission. The UN Security Council approved a plan for a 5,550-member force in late September, with the first foreign troops arriving in April. However, the full contingent has yet to arrive, and only an undisclosed number of troops from Chad have been deployed so far.

The Deadly Israeli Drone Strikes Three Israeli drone strikes on cars on a major highway linking Beirut to southern Lebanon have killed at least eight people, including two children, Lebanon’s Ministry of Health reported. A photograph of the bombed cars shared by Lebanon’s National News Agency following the attacks on Wednesday in the Jiyeh area, some 20km (12 miles) south of the Lebanese capital, showed the vehicles severely damaged, their exteriors charred and torn apart. Escalating Conflict and Humanitarian Crisis Al Jazeera’s Zeina Khodr, reporting from Tyre in southern Lebanon, said the “conflict is only escalating”. “It is a conflict that is taking a high toll on the civilians who live in these areas,” she said. Lebanon and Israel are expected to hold a new round of direct negotiations in Washington on Thursday, brokered by the United States. Hezbollah, which has been launching attacks on northern Israel and on Israeli troops who have entered and occupied a section of southern Lebanon, says it opposes the negotiations in the US. The Humanitarian Toll On Wednesday morning, the Israeli military issued forced displacement orders for the residents of Meiss el-Jabal, Yanouh, Burj Shemali, Hula, Debl and Aabbasiyyeh, warning that it will soon act against these six southern Lebanese villages “forcefully”. Anyone who remains “endangers their life,” the military said, warning residents to move at least 1,000 metres (0.6 miles) away to “open areas”. After this new round of forced displacement orders – which have been happening almost daily in the past week – Al Jazeera’s Khodr said one of the few remaining hospitals in the area was in the displacement zone. At least 100,000 people still live in the district of Tyre. 13 people were killed in attacks on towns in the south on Tuesday. Two Lebanese Civil Defence paramedics were among the dead. At least 380 people have been killed during the truce. The total death toll since the Israeli invasion and bombardment began on March 2 is more than 2,800. 108 emergency medical services and healthcare workers have been killed in Lebanon during the war. The Future Outlook “All of this is having a huge impact here on the communities in southern Lebanon,” Al Jazeera’s Obaida Hitto said from Tyre. “And there is a growing humanitarian crisis, with over a million people displaced.”

President Vladimir Putin announced on May 13, 2026 that Russia’s new Sarmat ICBM test was successful, branding it the most powerful missile ever built and signalling a major step in Moscow’s nuclear modernisation.Putin Announces Successful Sarmat Test LaunchState TV showed Sergei Karakayev, commander of Russia’s strategic missile forces, briefing the president on the test conducted on Tuesday. The Sarmat, dubbed “Satan II” in the West, is slated to enter combat service before the end of the year.Technical Specs and Performance ClaimsRange: exceeds 35,000 km (about 21,750 miles) via sub‑orbital flight.Warhead yield: claimed to be more than four times that of any current Western ICBM.Replacement goal: to supplant roughly 40 aging Soviet‑era Voyevoda missiles with higher precision.Development timeline: program started in 2011; prior to this test only one successful launch and a 2024 catastrophic failure were recorded.Strategic Implications for Global Arms ControlThe test occurs against the backdrop of the New START treaty’s expiration in February 2026, leaving the United States and Russia without a binding cap on strategic warheads. Both sides accuse each other of non‑compliance, and no successor agreement is in sight, raising concerns about a new arms‑control vacuum.U.S. officials, including former President Donald Trump, have floated the idea of a trilateral treaty that would also involve China, whose nuclear arsenal, while smaller, is expanding.Potential Trajectory of Russia’s Nuclear ModernisationRecent additions: Avangard hypersonic glide vehicle (already in service), Oreshnik IRBM (used in Ukraine), Poseidon underwater drone (final development stage), Burevestnik nuclear‑powered cruise missile.Strategic rationale: counter perceived U.S. missile‑defence shield and ensure second‑strike capability.Putin framed these developments as a response to a “new reality” where maintaining strategic parity is essential for Russia’s security.Outlook: Risks and Possible Diplomatic PathsAnalysts warn that the Sarmat’s deployment could accelerate a new arms race, especially if the United States expands its own missile‑defence and offensive capabilities. However, the urgency of re‑engaging in arms‑control talks may grow, as the lack of a treaty increases the risk of miscalculation.Future scenarios range from renewed high‑level dialogue leading to a multilateral framework that includes China, to a continued escalation where each side expands its nuclear arsenal to offset perceived vulnerabilities.

ACCESS: Medicare’s First AI‑Enabled Outcome‑Based Payment Model Pair Team was announced on April 30 as one of 150 organizations accepted into ACCESS (Advancing Chronic Care with Effective, Scalable Solutions), a CMS initiative that launches on July 5. The program shifts reimbursement from traditional time‑based fees to payments tied to measurable health outcomes such as lower blood pressure or reduced pain, covering conditions like diabetes, hypertension, chronic kidney disease, obesity, depression, and anxiety. Revenue Scale and Funding Behind Pair Team Staff: roughly 850 clinical professionals, the largest community‑health workforce in California. Revenue: exceeds nine figures (>$100 million) annually. Capital raised: about $30 million from investors including Kleiner Perkins, Kraft Ventures, and Next Ventures. Patient reach: partnerships give access to ~500,000 potential patients, with a goal of 1 million within three years. Industry context: digital‑health funding hit its highest Q1 total since the pandemic, with AI firms capturing the bulk of new capital. How Outcome‑Based Payments Could Redefine Chronic Care Delivery The ACCESS model creates the first federal mechanism to pay for AI agents that monitor patients between visits, coordinate social services, and ensure medication adherence. Flora, Pair Team’s voice‑AI assistant, now handles 24/7 intake, referrals, and check‑ins, delivering hour‑long conversations that act as both clinical touchpoints and companionship for high‑needs patients. Peer‑reviewed research in the Journal of General Internal Medicine shows Pair Team’s community‑integrated approach cuts avoidable emergency and inpatient utilization, with one‑in‑four hospital visits and one‑in‑two ER visits averted for its members. Risks remain: the program funnels highly sensitive data into a federal system with a history of breaches, and past CMS innovation pilots have drawn criticism for increasing federal spending without delivering projected savings. What’s Next for AI‑First Health Providers Under ACCESS Batlivala argues that lower per‑patient reimbursement rates are intentional, forcing providers to adopt lean, AI‑driven operations. As the program scales, success will hinge on: Automating patient interactions to keep costs below payment thresholds. Demonstrating measurable outcome improvements across the covered chronic conditions. Managing data‑privacy concerns to maintain trust among vulnerable populations. Attracting additional capital as investors watch the first AI‑centric Medicare payment model unfold. If Pair Team and its peers can prove the model’s efficacy, ACCESS could become a template for nationwide AI‑enabled, outcome‑based reimbursement, reshaping how Medicare incentivizes technology in health care.