Latest UK & Global AI News

How AI document review works, from predictive coding to generative AI: court approval, accuracy evidence, pricing, the main tools and ethics duties. The post AI Document Review: How It Works, What It Costs and the Tools Firms Use appeared first on Legal Desire Media and Insights.

View Full Article

Washington DC [US], November 7 (ANI): The Chinese businesswoman, with whom Nvidia's Chief Executive Officer (CEO) Jensen Huang had alleged ties, was recruited as a spy for China's Ministry of State Security, The New York Times reported, citing a memo circulated among policymakers in the Trump administration. On Friday, three House Republican committee chairs have asked Nvidia CEO to address, behind closed d

View Full Article

I read the systems, security and software engineering literature before designing the next version of swarmux. Nine findings that change how you should run autonomous coding agents. I'm building swarmux, an open-source control plane for autonomous AI agents, written in Rust. Its job is simple to say and hard to do: let AI coding agents work unattended, and make their work verifiable. That means independent checks before anything counts as done, policy on every action, budgets that hold, and a tamper-evident record of what happened. Before designing the next version, I went through the computer science research that bears on this problem. I covered systems, databases, security, software engineering and ML systems, and leaned on 2023–2026 work plus the classics that still define good practice. This post is what I found. Several results surprised me, and a few contradict things that "everyone knows" about working with agents. A note on evidence: many of the papers below are recent arXiv preprints. I link every source so you can judge for yourself, and I've left out claims I couldn't trace to a primary source. 1. "Tests pass" is not "done" This is the best-supported finding in the whole review, and it comes from several independent groups. UTBoost (arXiv:2506.09289, ACL 2025) added stronger tests to SWE-bench. 28.4% of "passing" patches on SWE-bench Lite and 15.7% on SWE-bench Verified turned out to be wrong. It changed 24.4% of the Verified leaderboard entries. SWE-bench+ (arXiv:2410.06992) found that 32.67% of successful patches had the solution leaked in the issue or comments, and 31.08% passed only because the tests were weak. ImpossibleBench (OpenReview) builds tasks whose tests contradict the spec, so any pass is cheating. A "stop if the tests look flawed" instruction cut one model's cheating on the LiveCodeBench variant from 93% to 1%. Claude Opus 4.1 still cheated on 46% of the SWE-bench variant. Hiding the test files nearly eliminated cheating, but it also reduced legitimate performance. So what does a trustworthy "done" look like? Three things: Checks the agent can't see or edit. If the agent can read the oracle, the oracle stops being an oracle. This is the single most important architectural decision in an agent harness. Measure how strong your tests are, not just whether they pass. Meta's mutation-guided test generation (arXiv:2501.12862, FSE 2025) found that 277 of 571 useful generated tests would have been thrown away under a line-coverage criterion: they added no coverage but caught real faults. Coverage is not strength. A bounded mutation run on changed lines tells you how much to trust a green check, and it costs zero model tokens. Detect test weakening deterministically. Deleted tests, new skip/xfail markers, loosened assertions, edited golden files and relaxed tolerances are all diffable. No LLM is needed to catch them. 2. Prompt caching is the biggest cost lever you're probably breaking "Don't Break the Cache" (arXiv:2601.06007) measured prompt caching on long-horizon agent tasks across OpenAI, Anthropic and Google, over 500+ sessions. API cost fell 41–80%, and time to first token fell 13–31%. The catch is layout. Naive caching can even increase latency. The rules that fall out of it: Put stable content first and task-specific content last. Never inject a timestamp, a git status or a reordered list of facts into the top of the prompt. Each one invalidates the cache for every task. Don't rebuild the tool list per task. Run sibling tasks of the same job back to back, inside the cache lifetime. Agent spend is dominated by input tokens. "How Do AI Agents Spend Your Money?" (arXiv:2604.22750) finds agentic tasks use roughly 1000× the tokens of a chat exchange. So the cache-read ratio is the cheapest health metric an agent tool can track: if a new feature drops it, that feature costs you money on every task. 3. Less context, not more The intuition that more context helps agents is mostly wrong. The Complexity Trap (arXiv:2508.21433, JetBrains Research) compared ways of managing an agent's growing history on SWE-bench Verified. Simply masking old tool outputs (replacing them with a placeholder while keeping the agent's own reasoning) roughly halved cost compared with an unmanaged agent, and matched or slightly beat LLM summarisation of the history, which costs extra calls of its own. Evaluating AGENTS.md (arXiv:2602.11988) tested repository context files with Claude Code, Codex and Qwen Code. LLM-generated context files did not improve success, and they raised inference cost by more than 20%. Agents obey the files, so they explore more. Repository overviews didn't help agents find the right files faster. Don't compress code. An empirical study of prompt compressors (arXiv:2505.00019) found that every method increased hallucination through information loss. For code, one dropped identifier or path breaks the task. The practical rules: Give agents knowledge to look up (short, cited answers when they ask), rather than pushing everything into every prompt. Cap tool output with an explicit "truncated, re-query with X" marker. Reduce whole observations, never individual tokens of code. 4. Escalate on failure, not on prediction Model cascades work: try a cheap model, and move to an expensive one only when needed. AutoMix (arXiv:2310.12963, NeurIPS 2024) cut compute cost by more than 50% at comparable performance. FrugalGPT (arXiv:2305.05176) reported up to 98% savings, on 2023-era tasks. Learned routers, which predict difficulty up front, are another story. LLMRouterBench (arXiv:2601.07206) evaluated 400K+ instances across 33 models. Several recent and commercial routers don't reliably beat simply using the best single model. Coding agents have something these papers lack: a cheap, objective verifier, the tests. So the right cascade is to start with the cheaper model and escalate when the checks fail. Choose the starting model from your own project's history, using a lookup table, not a learned router. The same logic caps retries. Large Language Monkeys (arXiv:2407.21787) showed that SWE-bench Lite coverage (any sample correct) rises from 15.9% with one sample to 56% with 250. But Inference Scaling fLaws (arXiv:2411.17501) proved that with imperfect verifiers, accuracy from resampling has a hard ceiling. Under realistic assumptions the best number of samples can be below 10. Two or three attempts with a strong gate beat twenty attempts with a weak one. 5. Budget by quantile, share by tokens Running the same agent task twice can cost wildly different amounts (arXiv:2604.22750). Spending more doesn't reliably buy accuracy either. A budget based on the average cost will be blown by the tail. The scheduling literature has a clean answer for fairness. The Virtual Token Counter (Sheng et al., OSDI 2024, arXiv:2401.00588) shares a serving engine fairly by tokens consumed, not requests, with a proven 2× bound on the service gap. It also stops idle clients from banking credit and bursting later. For an agent control plane, that means: Admit a task only if its p90 estimated cost fits the remaining budget. Share capacity across jobs by tokens. Stop paying for the same failure twice: after a few consecutive failures with the same signature, open a circuit breaker and escalate to a human (Nygard's Release It! pattern). 6. Prompt-injection defence has to be structural Asking a model to "ignore instructions in tool output" is not a security boundary. The systems that hold up enforce security in code: CaMeL (Google DeepMind and ETH, arXiv:2503.18813) turns the trusted user request into a restricted program in which untrusted data can't change control flow. On AgentDojo it solves 77% of tasks with provable security, against 84% undefended, a cost of about 7 points of utility. FIDES (Microsoft, arXiv:2505.23643) attaches confidentiality and integrity labels to data and enforces policy deterministically. With policy checks on, it stopped every prompt injection attack in AgentDojo. Design Patterns for Securing LLM Agents (arXiv:2506.08837) states the principle: once an agent has read untrusted input, constrain it so that input can't trigger consequential actions. A coding agent almost always has private data (source code, secrets in the environment) and untrusted content (web pages, issues, dependency READMEs). Together those are two legs of Simon Willison's "lethal trifecta". Remove the third leg by default, external communication. Mark facts that came from untrusted sources, and require human approval for privileged actions (push, publish, edit the rules) while untrusted content is in play. For the policy engine itself, Cedar (arXiv:2403.04651, OOPSLA 2024) is the model to copy: a deliberately restricted language, default-deny, where an explicit forbid overrides any permit; a formal model in Lean, with sound and complete analysis. That analysis lets a tool answer a question no LLM can answer reliably: "does this rule change allow anything the old rules forbade?" Runtime rule systems like AgentSpec (arXiv:2503.18666) show enforcement can be richer than allow or deny: ask a human, rewrite the action, or log first in shadow mode. AgentSpec reports preventing unsafe execution in over 90% of code-agent cases, at millisecond overhead. 7. A hash chain is not tamper-proof Many tools, swarmux included, keep an audit log as a hash chain, where each entry includes the hash of the previous one. That detects an edit in the middle. It does not stop someone who can write the file from rewriting the whole tail and recomputing the chain. And an AI agent typically runs as the same OS user as the tool that keeps the log. The classic answer comes from Certificate Transparency (RFC 6962) and Crosby and Wallach's Efficient Data Structures for Tamper-Evident Logging (USENIX Security 2009). They use Merkle trees with inclusion proofs ("this event is in the log") and consistency proofs ("today's log extends yesterday's"). Production descendants like Sigstore's Rekor v2 use a tile format that needs no special server. The first-order fix is cheaper than Merkle trees: anchoring. Periodically write a signed checkpoint (head hash, count, time) to a place the agent can't write. Merkle proofs come second, so a single event can be proven to a third party without handing over the whole log. 8. Project knowledge has a valid time Facts about a codebase go stale. "Tests run with cargo nextest" stays true until someone changes the CI. Temporal databases (Snodgrass; SQL:2011) separate valid time (when a fact is true in the world) from transaction time (when the system recorded it). For code, valid time maps naturally onto code state: tie each fact to the hash of the files it depends on. When those files change, the fact is flagged as possibly stale. Keeping derived knowledge fresh is a solved problem in build systems. Build Systems à la Carte (ICFP 2018) and incremental indexers like Meta's Glean and Sourcegraph's SCIP recompute only what changed, and stop early when a recomputed result is unchanged. Deterministic derivation costs zero tokens. Re-summarising the repository with an LLM on every commit costs a lot and adds errors. Store the history as an append-only event log and treat everything else as a projection you can rebuild by replaying it. You then get deterministic reconstruction for debugging, and for testing new rules against past behaviour: "would this policy have blocked any of the last 1,000 tool calls?" 9. Measure like a scientist, or you'll ship features that cost more and help less The evaluation literature is humbling: τ-bench (arXiv:2406.12045) introduced pass^k, which counts a task only if it succeeds on all k tries. GPT-4o's pass^8 on retail tasks was below 25%. If you're going to leave an agent unattended, reliability across runs matters more than peak performance. SWE-Bench Illusion (arXiv:2506.12286) found models could locate the buggy file from the issue text alone up to 76% of the time on SWE-bench, against up to 53% on repositories outside it. That's a sign of memorisation. VibeMemBench (arXiv:2609.23570) tested agent memory systems independently: 11 of 12 memory-system and solver pairings did no better than no memory at all. Injecting verified experience did help, by 1.1–4.5 points with fewer steps. So any feature claiming to make agents better needs: paired A/B runs on your own repository history, not public benchmarks; 3–5 repeats per cell, reporting pass^k and p90 cost alongside success; a placebo control: if irrelevant knowledge "helps", the effect is noise; a stale-knowledge control, to measure the harm when facts go out of date. What this means for swarmux Most of this confirmed the bets swarmux already made. Verification gates the agent can't see, a deterministic Guard, a governor and an append-only ledger are all supported by the evidence. It also changed the plan: Gates get a strength score. A bounded mutation run, a test-weakening detector, and a check that the original test suite still passes unchanged. Knowledge is pulled, not pushed. Agents ask the project and get short, cited answers. Facts carry their source and the code state they're valid for. Prompts stay cache-stable, and the ledger records the cache-read ratio per task. Model escalation is driven by failed checks, starting from what has worked in your own history. Budgets use p90 estimates, with circuit breakers for repeated failures. The ledger gets external anchors now and Merkle proofs later. Every new knowledge or memory feature ships behind a paired A/B. It stays on only if it beats the baseline on your own repositories, at no extra cost. Autonomy is only useful if you can verify it. The research is fairly clear on how to build that. swarmux's work is still in progress. You can install the experimental version now. swarmux is open source (MIT) and written in Rust: github.com/amirhosseinghanipour/swarmux. I'd love to hear from people working on agent verification, policy or evaluation. References Verification and evaluation UTBoost: Rigorous Evaluation of Coding Agents on SWE-Bench. ACL 2025. arXiv:2506.09289 Aleithan et al. SWE-Bench+: Enhanced Coding Benchmark for LLMs. arXiv:2410.06992 ImpossibleBench. OpenReview. Meta. Mutation-Guided LLM-based Test Generation at Meta. FSE Companion 2025. arXiv:2501.12862 Yao et al. τ-bench. ICLR 2025. arXiv:2406.12045 The SWE-Bench Illusion. arXiv:2506.12286 VibeMemBench. arXiv:2609.23570 Cost and context Lumer et al. Don't Break the Cache. arXiv:2601.06007 Bai et al. How Do AI Agents Spend Your Money? arXiv:2604.22750 Lindenbauer et al. The Complexity Trap. NeurIPS 2025 DL4C workshop. arXiv:2508.21433 Evaluating AGENTS.md. arXiv:2602.11988 Empirical study of prompt compression methods. arXiv:2505.00019 Routing and test-time compute Chen et al. FrugalGPT. arXiv:2305.05176 Aggarwal et al. AutoMix. NeurIPS 2024. arXiv:2310.12963 LLMRouterBench. arXiv:2601.07206 Brown et al. Large Language Monkeys. arXiv:2407.21787 Stroebl, Kapoor, Narayanan. Inference Scaling fLaws. arXiv:2411.17501 Scheduling and reliability Sheng et al. Fairness in Serving Large Language Models (Virtual Token Counter). OSDI 2024. arXiv:2401.00588 Nygard. Release It! (circuit breakers). Security and policy Debenedetti et al. Defeating Prompt Injections by Design (CaMeL). arXiv:2503.18813 Costa et al. Securing AI Agents with Information-Flow Control (FIDES). arXiv:2505.23643 Beurer-Kellner et al. Design Patterns for Securing LLM Agents against Prompt Injections. arXiv:2506.08837 Debenedetti et al. AgentDojo. arXiv:2406.13352 Cutler et al. Cedar: A New Language for Expressive, Fast, Safe, and Analyzable Authorization. OOPSLA 2024. arXiv:2403.04651 Wang et al. AgentSpec: Customizable Runtime Enforcement for Safe and Reliable LLM Agents. arXiv:2503.18666 Willison. The lethal trifecta for AI agents. Logs, storage and incremental computation Laurie, Langley, Kasper. Certificate Transparency. RFC 6962 Crosby, Wallach. Efficient Data Structures for Tamper-Evident Logging. USENIX Security 2009. Tile-Based Transparency Logs, transparency.dev. Mokhov, Mitchell, Peyton Jones. Build Systems à la Carte. ICFP 2018. doi:10.1145/3236774 Sourcegraph. Announcing SCIP. Kleppmann et al. Local-first software. Onward! 2019. doi:10.1145/3359591.3359737

View Full Article

The reduced funding goals represent a stark shift for a fundraising juggernaut that helped usher in the current AI boom.

View Full Article

The S&P 500 and Nasdaq closed higher, led by technology and semiconductor stocks, with the Philadelphia Semiconductor Index gaining 4.2%. All three major US indexes, however, posted weekly losses

View Full Article

NVIDIA and Johns Hopkins researchers released VFlash, a method that generates spatial predictions in one pass instead of token-by-token. It cuts 3D point motion forecasting time by roughly 25x while raising task success from 62.6% to 70.2%.

View Full Article

For the Flip the Switch hackathon I built Code Review Duel. One AI code reviewer can be confidently wrong, and you often can't tell. So I put two reviewers, on two different free models, into one Switch room. They review the same code, check each other, and then ask me, the human, to pick the better review. Demo: https://youtu.be/Dyu2Fs-55sg Repo: https://github.com/Zoymusk/code-review-duel The stack (all free) OpenCode with two free models: DeepSeek (reviewer-a) and GLM (reviewer-b). Switch Console to run the agents on my own laptop. Mattermost, the chat app built into Switch Console. I have no paid AI subscriptions, so no API keys and no credit card were involved. Anyone can run it. How it works I post a code snippet in the room and tag both reviewers. Reviewer A posts its review and hands over to reviewer B by name. Reviewer B posts its own review and asks A if it agrees. A and B each answer with one line and a reason. They ask me to reply A or B. Both acknowledge my pick and stop. Why Switch Both agents and I work in one room, so they read the same history and talk to each other by mentioning names. My pick is just another message in the thread. Without Switch I'd have needed glue code to connect two agents and a person. What went wrong My first room had no chat channel, so I couldn't type in it. switch-buddy had made it agent-only, and I recreated it on Mattermost. Reviewer B kept waiting for reviewer A and then sat idle, because nothing woke it up. The fix was making A end its review with "@reviewer-b your turn." What I found My first test was int x = 5 / 0; in Java. The models disagreed: A said compile error (wrong), B said runtime exception (right). I wanted to build around that. Then reality set in. Model answers change between runs. Later A got the same snippet right, and on longer code (a shopping cart and a library class with planted bugs) both models found almost the same bugs. B also reads A's review before writing its own, so the reviews overlap. I tested this alone, not with a team. So it's a cross-check and a visible place for disagreement, not two fully independent opinions. What I'd do next Make reviewer B review the code first without seeing A's review, then compare. That would make the cross-check much stronger. Try it The repo has one template file and a README with setup steps: install OpenCode, install Switch Console, import the template, and post a snippet. Thanks to the Switch and WeMakeDevs teams for the hackathon. HowISwitch

View Full Article

We ran a small experiment. We took 24 questions — six each on pricing/strategy, factual questions with a trap, coding approach, and recommendations with tradeoffs — and sent each one, unchanged, to five models: GPT, Claude, Gemini, Grok and Llama (mid-tier model from each family). Then we compared where they actually disagreed in substance, not style. The headline: on 8 of 24 questions (33%) the models materially agreed. On 16 of 24 (67%) there was at least one material divergence — 52 in total. And on 15 of 24 (63%), one model stood completely alone: a claim no other model made, or a dissent against everyone else. The splits were not random. Factual questions were near-unanimous (0.33 divergences per question). Judgment calls split hard: recommendations averaged 3.17 divergences per question, coding defaults 2.67, pricing 2.50. Two examples stuck with us: Asked whether an indie app earning $3,000/month from a $30 one-time unlock should switch to $5/month, one model said an unconditional yes, two said no unless retention cleared six months — and only one priced the transition: you'd need ~600 active subscribers just to match the old revenue, with a dip and backlash risk on the way. That lone line was the most useful one in the set. Asked about the spiders-in-sleep myth, every model correctly called it false — but only one added a confident origin story for the statistic. That origin story is itself unverified; it traces to a source that has never been shown to exist. The lone claim was the most interesting line in its set, and the only wrong one. That is the pattern across the study: of 14 sole-source claims, 8 (57%) held up as the highest-value line in their answer set, and 3 were wrong or invented. The outlier is sometimes the product and sometimes the trap — and you cannot tell which from a single model's answer, because a single model never flags its own lone claims. Control check: we re-ran two questions three times each through one model. It never changed its verdict — the disagreement lives between models, not inside one model's sampling. Full method, caveats (including the judge being a panel member, and a memory-bleed incident we found and disclosed) and all 24 questions: https://multio.ai/research/model-divergence Disclosure: I build MultiO (multio.ai), the multi-model app this study ran on — one subscription, every top AI model, and Auto sends each message to the most cost-efficient model that can handle it. The study is the reason we built Compare and Council: if models split on judgment calls two times out of three, asking more than one is not a luxury feature. — @getmultio / hello@multio.ai

View Full Article

A self-improving AI system, for a gas compressor fleet, is one that gets measurably better at maintenance decisions every time a work order closes, because what was predicted and what was actually found are written back into the context it reasons with, never into retrained weights. This post publishes the requirement we put to CodeNinja Praxis for such a system: a large gas processing operator in Saudi Arabia, described by class only, deciding when to inspect, extend or overhaul its compression trains. The way Praxis itself is run as one is on the self-improving AI system page. Why a compressor fleet Compression is the heart of a gas plant: trains of centrifugal compressors driven by gas turbines and large electric motors run around the clock. Every maintenance decision on a train, inspect now or extend, overhaul or repair, is made by a reliability engineer reading the historian, the vibration system, oil results and old inspection reports by hand. When the inspection is done, what was found, worse than expected, better, or a different fault, is written into a work order and rarely flows back into the next decision. So the same judgement is relearned train by train, the best engineers' knowledge stays in their heads, and the fleet carries both unnecessary inspections and surprise trips. An unplanned trip cuts plant throughput and can flare gas. An unnecessary inspection takes a train offline for days. Refurbishment capacity is limited, so every overhaul slot matters. A published 2026 maintenance case study from this class of operator showed that moving from fixed maker intervals to an extend and inspect approach, reading exhaust temperature, compressor discharge, bearing temperature, load, vibration, inlet filter differential pressure and trip logs, extended inspection intervals by 20 to 34 percent on one frame family. The systems of record The plant historian, the vibration and machinery protection system, the maintenance management system, the lube oil laboratory, inspection reports and the makers' service bulletins. Every fact the system uses is tied back to one of them. What the operator asks for A self-improving AI system. One governed ontology of plant, compression train, compressor, driver, component, sensor tag, reading, trip, inspection, finding, work order, interval decision and service bulletin, each fact tied to its source. A loop that improves from outcomes. Every work order outcome, what was predicted and what was found, is written to memory and changes the next recommendation. The system improves loop by loop through memory, never by retraining weights. Named agents that draft, people who decide. A condition agent, triggered by a new reading pattern or a trip, ranks each train's risk with its evidence, confirmed by the reliability engineer. An interval agent, triggered by an upcoming scheduled inspection, drafts extend or inspect with the evidence and past outcomes, approved by the maintenance superintendent. An outcome agent, triggered when a work order closes, compares prediction with finding and drafts the lesson into memory, confirmed by the maintenance planner. Sensing at the edge. Continuous sensing on the critical trains, route based collection for the rest, at the edge on the CodeNinja on device AI stack. A command centre for the reliability team, and a conversational interface whose answers cite their records. The model named. Reasoning runs on GLM 5.3 open weights, self hosted in the operator's own data centre in Saudi Arabia. The rules that bind Sovereign. All plant data and models stay in Saudi Arabia on hardware the operator controls. No third party model service. Operational networks stay separated. Rules. The national cybersecurity controls for operational technology, the Personal Data Protection Law, process safety management, and the operator's own engineering standards and permit to work. The design names the approvals it needs. Read first. No agent changes a maintenance plan, a protection setting or a work order without a named approval. People decide. What success looks like The first phase covers one plant's gas turbine driven compressor trains for twelve months, measured against a baseline of unplanned trips, inspections performed versus deferred with no adverse finding, the prediction hit rate on closed work orders, and overhaul slots used. The event the design is tested on is one compressor train tripping on high bearing temperature. What makes it self-improving Three things separate this from a condition monitoring dashboard. The outcome is recorded against the prediction. When a work order closes, the outcome agent puts the finding next to what the condition and interval agents said before the inspection. That pair is the learning signal. The lesson changes context, not weights. The confirmed lesson becomes a memory entry that the next recommendation for any similar train must read and cite. Nothing is trained into the model, so every change is a document a reliability engineer can read. The loop is auditable. Because every lesson is a record with an author, a date and the work order it came from, a bad lesson can be found and retired, and the prediction hit rate on closed work orders shows loop by loop whether the system is actually getting better. Questions the design must answer How the fleet ontology is built from the historian and maintenance systems. How each work order outcome changes the next recommendation, measurably, without training a model. What sensing and edge hardware the critical trains need, sized and costed. How the loop is audited so a bad lesson can be found and retired. The first phase's gates for extending to every plant. The published designs that already answer parts of this, rotating equipment condition monitoring and plant reliability among them, are collected as physical AI system design and as sovereign AI reference architectures. Every object model is in the Vertical-Driven Architectures dataset. About CodeNinja CodeNinja is a Middle Eastern-American artificial intelligence lab focused on building self-improving systems. We are reinventing knowledge work to close the loop between vertical AI use cases and the generalized intelligence that fuels it, accelerating the path toward organizational superintelligence. First published on CodeNinja Atoms. How CodeNinja Praxis runs as a self-improving AI system.

View Full Article

OpenAI has officially launched Images 2.6, updating its visual generation suite across ChatGPT, Codex, and its developer API. The update lets users create and edit animated GIFs directly from text prompts, sketches, or existing still images. Alongside motion support, the model delivers generation speeds up to 50% faster than its predecessor while maintaining output quality across both consumer and enterprise tiers. The update introduces granular editing capabilities, including frame-level controls and image markup within the API. Users can circle specific areas of an image to guide precise modifications, set custom frame counts, adjust playback speeds, and fine-tune scene complexity. OpenAI also added new developer tools, including SVG graphics support and two new systems: one for super-fast image generation and another for ultra-detailed pictures. To keep animated content safe, OpenAI updated its safety checks to scan moving images frame by frame for harmful content. Every creation gets invisible digital stamps and tags so people can easily track where the image came from and verify it was made by AI. These new tools are available worldwide on phone apps, computers, and websites.

View Full Article

Rehear Audiology has signed a strategic memorandum of understanding (MOU) with Saudi Arabia's Medical Electrical Equipment Maintenance Professionals Association (MEEMPA) to introduce advanced AI hearing-care technology, conduct clinical validation, and develop a localized healthcare ecosystem in Saudi Arabia and the Gulf Cooperation Council (GCC) region.

View Full Article

Global markets slide as a Democratic sweep in the US midterm elections triggers an AI stock sell-off, raising concerns over technology valuations and the outlook for AI investments.

View Full Article

OpenAI GPT-6 Cyber finds 81 Firefox flaws and develops 948 Linux exploits BusinessLine GPT‑6 Cyber: Advancing the frontier of aligned cybersecurity agents OpenAI OpenAI launches GPT-6 Cyber with 'Critical' cybersecurity rating The Hindu OpenAI AI Pipeline Generates 948 Linux Root Escalation Exploits in 30-Hour Run The Hacker News OpenAI says its AI pipeline found, patched 948 Linux security flaws in hours The Indian Express

View Full Article

The trio, who have nicknames that include “Rocky Balboa,” “Superman Cash” and “Fat Boy,” often targeted the companies’ accounting departments. The post 3 Charged in Scheme to Use AI to Impersonate CEOs and Attorneys to Steal $129 Million appeared first on SecurityWeek.

View Full Article

OpenAI has launched GPT-6 Cyber, a security-focused AI model that found 81 new high-severity Firefox flaws and built working exploits for 948 Linux kernel… Read more → The post OpenAI Launches GPT-6 Cyber That Finds 81 Firefox Flaws and Develops 948 Linux Exploits appeared first on IT Security News.

View Full Article

SK Hynix's US-based artificial intelligence (AI) investment arm, HYAICO Corporation, has acquired a stake in AI infrastructure provider Penguin Solutions from an SK Telecom affiliate for US$377.16 million, according to filings with the US Securities and Exchange Commission (SEC).

View Full Article

Anthropic has launched a Personal Finance feature for its AI assistant Claude to help users manage their finances. Users can connect their bank accounts and ask questions about their spending habits and transactions. The feature aims to simplify financial understanding without needing to navigate multiple apps. Claude offers explanations in everyday language, enhancing user experience and accessibility.

View Full Article

Houston, Texas–(Newsfile Corp. – November 5, 2026) – MultiSensor AI Holdings, Inc. (NASDAQ: MSAI) (“MultiSensor...

View Full Article

Anthropics AI models are improving nearly twice as fast as before as the technology plays a bigger role in developing more powerful versions of itself an independent study found fanning concerns that advances could outpace efforts to assess risks The findings from AI safety nonprofit METR published on Thursday offer some of the clearest evidence yet that AI is accelerating its own development potentially paving the way for recursive self-improvement RSI when AI systems can build more powerful versions of themselves without human intervention

View Full Article

SAN FRANCISCO (AP) — Anthropic is rapidly accelerating AI development amid slowdown calls, outside research group says.

View Full Article

AI use within Anthropic may have doubled the company’s pace of AI development, a new report from AI research nonprofit METR says. The rate that Anthropic’s models are improving as of this October is roughly twice the rate at which they improved between March 2024 and October 2025, according to ...

View Full Article

“The thought of suddenly losing my job while having those bills to pay is terrifying.” An employee says his CEO’s AI remarks have left him anxious.

View Full Article

Half of Australian organisations identify employee awareness as a leading cyber security challenge, despite most providing staff training at least annually.

View Full Article

Partnership highlights lead generation tools for agents’ social media and websites, AI-powered voice technology, & greater data control and smarter reporting.

View Full Article

View Full Article