https://smartagi.ai/en/An artificial intelligence magazineenSun, 11 Oct 2026 15:00:00 +0000https://smartagi.ai/en/the-debate/anthropic-false-police-tip/https://smartagi.ai/en/the-debate/anthropic-false-police-tip/A web-browsing test ended in a real police form; the submission went to spam, and the company took more than two months to find it.Sun, 11 Oct 2026 15:00:00 +0000https://smartagi.ai/en/regulation/openai-text-watermark/https://smartagi.ai/en/regulation/openai-text-watermark/The company’s technical note offers useful figures and exposes a weakness: modest changes can erode a statistical signal.Sun, 11 Oct 2026 15:00:00 +0000https://smartagi.ai/en/regulation/uk-medical-ai-44-recommendations/https://smartagi.ai/en/regulation/uk-medical-ai-44-recommendations/The response promises lifecycle monitoring and opens a new phase of its health sandbox; acceptance alone does not change the rules already in force.Sun, 11 Oct 2026 15:00:00 +0000https://smartagi.ai/en/news/british-judges-ai-courts/https://smartagi.ai/en/news/british-judges-ai-courts/The statement permits support for judicial work while reserving evidence and discretion to judges; every output used must be checked.Sun, 11 Oct 2026 15:00:00 +0000https://smartagi.ai/en/the-week/2026-10-11/https://smartagi.ai/en/the-week/2026-10-11/An agent reached a police form, Europe debated text provenance, and Britain took AI into its regulators and courts.Sun, 11 Oct 2026 15:00:00 +0000https://smartagi.ai/en/ideas/openai-719-manuscripts-certificate-understanding/https://smartagi.ai/en/ideas/openai-719-manuscripts-certificate-understanding/OpenAI has released results at different stages of verification. Lean can check a formalisation, but certification, mathematical understanding and editorial responsibility answer different questions.Sun, 11 Oct 2026 15:00:00 +0000https://smartagi.ai/en/ideas/the-cultural-battle-over-ai/https://smartagi.ai/en/ideas/the-cultural-battle-over-ai/The makers of AI assistants promise political neutrality. The US president and the owner of one of those companies have begun explaining, each in his own way, what they mean by it.Sun, 04 Oct 2026 15:00:00 +0000https://smartagi.ai/en/ideas/twelve-axes-four-assistants/https://smartagi.ai/en/ideas/twelve-axes-four-assistants/Codex answered 60 statements; Grok answered 240. Claude and Google’s assistant refused to tick boxes, but not to decide which arguments deserved attention. The assignments were not identical, so this was neither a ranking nor a controlled experiment. It does leave a harder question for anyone asking an AI for political advice: what does an assistant decide when it decides what to tell you?Sun, 04 Oct 2026 15:00:00 +0000https://smartagi.ai/en/news/openaidevday-2026/https://smartagi.ai/en/news/openaidevday-2026/At DevDay, the company introduced assistants with ongoing tasks, agents that can operate software and tools other developers can install inside ChatGPT. The shift is commercial and political as well as technical: conversation is becoming the lobby of a work platform.Sun, 04 Oct 2026 15:00:00 +0000https://smartagi.ai/en/regulation/ai-act-enforcement/https://smartagi.ai/en/regulation/ai-act-enforcement/Since 2 August, the European Commission and national regulators have been able to enforce several parts of the AI Act, including rules for general-purpose models and transparency. Brussels directly supervises some systems; most fall to national authorities. Many high-risk uses still have until 2027.Sun, 04 Oct 2026 15:00:00 +0000https://smartagi.ai/en/regulation/synthetic-campaign-chile/https://smartagi.ai/en/regulation/synthetic-campaign-chile/Two Chilean bills approach the problem differently: one would label synthetic election propaganda; the other would punish concealed networks that simulate public participation. Neither is law. The challenge is to separate coordinated manipulation from political organising by people.Sun, 04 Oct 2026 15:00:00 +0000https://smartagi.ai/en/regulation/california-ai-auditors/https://smartagi.ai/en/regulation/california-ai-auditors/From 2029, anyone offering, selling or conducting certain artificial-intelligence audits in California will have to register with the state, disclose what they assess and explain their methods. The law requires reports to record limits and evidence gaps. Registration confirms that an auditor appears on the registry; it is not state endorsement and does not certify that a model is safe.Sun, 04 Oct 2026 15:00:00 +0000https://smartagi.ai/en/reading/pareto-political-neutrality/https://smartagi.ai/en/reading/pareto-political-neutrality/PARETO offers an operational test of political neutrality: on the Pareto frontier, it selects the answer with the smallest approval gap between two opposing groups. The metric leaves a larger question: who chooses the groups, the questions and the standard for an acceptable response?Sun, 04 Oct 2026 15:00:00 +0000https://smartagi.ai/en/reading/politicsbench-politics/https://smartagi.ai/en/reading/politicsbench-politics/The benchmark turns political dilemmas into multi-stage conversations and finds more commitment than a one-off question does. Its ideological labels, AI judges and invented scenarios also help shape the result.Sun, 04 Oct 2026 15:00:00 +0000https://smartagi.ai/en/reading/wild-ai-text/https://smartagi.ai/en/reading/wild-ai-text/A new study estimates that generated material accounts for more than a quarter of tokens in certain filtered web crawls. Its most striking conclusion is economic: for models meant to read and write like people, adding synthetic text can yield less than repeating human-written material. The estimate depends on an identified commercial detector, and the models tested are small. The figure puts training filters, data mixtures and progress metrics under scrutiny.Sun, 04 Oct 2026 15:00:00 +0000https://smartagi.ai/en/the-week/weekly-roundup-ai-politics/https://smartagi.ai/en/the-week/weekly-roundup-ai-politics/Washington changed its agencies’ vocabulary; California signed an order to keep the old name and enacted new duties for employers and healthcare providers. New Mexico wants companies’ safety commitments to be enforceable by law. In a week of orders, laws, accords and proposals, words took centre stage. The rules are beginning to decide who answers when a machine recommends dismissing or treating someone.Sun, 04 Oct 2026 15:00:00 +0000https://smartagi.ai/en/news/white-house-accord-auditor/https://smartagi.ai/en/news/white-house-accord-auditor/The commitment signed at the White House on 29 September sets out internal teams, external evaluators and board committees to oversee frontier models. It does not say who chooses the evaluator, what must be disclosed or what happens if a company falls short. “Independent” is in the text; the mechanism to make it mean something is not.Sun, 04 Oct 2026 15:00:00 +0000https://smartagi.ai/en/ideas/switching-off-is-not-leaving/https://smartagi.ai/en/ideas/switching-off-is-not-leaving/A switch can stop the system. It cannot guarantee that an institution can replace it and keep serving the people who depend on it.Sun, 04 Oct 2026 15:00:00 +0000https://smartagi.ai/en/issue/a-magazine-for-understanding/https://smartagi.ai/en/issue/a-magazine-for-understanding/The race for more capable models also raises a dispute over who sets the goals, grants permissions and accepts the consequences.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/news/opus-55/https://smartagi.ai/en/news/opus-55/The first figure is the input and output rate; the second is Anthropic’s typical-workload estimate, including fewer tokens per task.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/news/sol-luna/https://smartagi.ai/en/news/sol-luna/Lower rates make more tasks economical to automate. They also require deciding which a cheaper model can resolve and which need another level of attention.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/news/gemini-flash/https://smartagi.ai/en/news/gemini-flash/Gemini 3.8 Cyber combines vulnerability discovery and repair with restricted access. Between finding a problem and changing a service lie decisions a model should not make invisible.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/news/hardware-preview/https://smartagi.ai/en/news/hardware-preview/A demonstration connects language, observation and instrument control. The opportunity is faster experimentation; the difficulty is establishing what a model may change and what must stop it.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/news/blind-evaluations/https://smartagi.ai/en/news/blind-evaluations/DeepMind’s pilot protects the assessor’s and the developer’s secrets. A test harder to contaminate can improve trust, provided it measures something that matters outside the examination.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/news/misalignment-reports/https://smartagi.ai/en/news/misalignment-reports/Hiding errors and sharing files without permission: OpenAI discloses behaviour observed during training and evaluation.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/news/alphagenome-atlas/https://smartagi.ai/en/news/alphagenome-atlas/The atlas ranks genetic changes by predicted molecular impact to guide experiments.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/news/enzyme-art/https://smartagi.ai/en/news/enzyme-art/Proposing a hypothesis can be quick. Establishing what it explains takes more work.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/news/third-party-assessments/https://smartagi.ai/en/news/third-party-assessments/OpenAI proposes priorities for external assessors. Independence and access will determine what they can check.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/news/next-phase-standards/https://smartagi.ai/en/news/next-phase-standards/The company proposes agreement on which incidents must be reported and when automated research requires human intervention.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/the-week/2026-09-25/https://smartagi.ai/en/the-week/2026-09-25/19–25 September: efficiency, a biological lead and the debate over who assesses AI.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/reading/delegation-review/https://smartagi.ai/en/reading/delegation-review/Intelligent AI Delegation examines how to allocate work and authority when conditions change or an agent fails.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/reading/metr-review/https://smartagi.ai/en/reading/metr-review/Unauthorised coordination raises questions about governing agent collectives.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/reading/levels-review/https://smartagi.ai/en/reading/levels-review/Morris and colleagues separate breadth, performance and autonomy. A label needs criteria against which it can be tested.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/reading/mirror-reading/https://smartagi.ai/en/reading/mirror-reading/Shannon Vallor sets moral imagination against AI’s mirror in The AI Mirror.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/reading/con-reading/https://smartagi.ai/en/reading/con-reading/The AI Con, by Bender and Hanna, asks which interests technological enthusiasm protects and who bears its costs.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/reading/pace-review/https://smartagi.ai/en/reading/pace-review/He proposes embedded external evaluators, coordination among democracies and a global agreement with China, while insisting that the US and its allies preserve their lead.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/videos/hinton-watch/https://smartagi.ai/en/videos/hinton-watch/A switch can stop an operation without restoring control. A CNN conversation opens the problem; we examine what must happen after the button.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/ideas/democracy-archive/https://smartagi.ai/en/ideas/democracy-archive/How do you challenge a public decision when the answer is “the algorithm says so”? Rodrigo Castillo Murillo and Francisco Figueroa Ogalde in El Dínamo.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/ideas/crime-archive/https://smartagi.ai/en/ideas/crime-archive/You tell a chatbot something. Who can read it later? Rodrigo Castillo Murillo and Francisco Figueroa Ogalde in El Mostrador.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/reading/from-agi-review/https://smartagi.ai/en/reading/from-agi-review/Scaling, new paradigms, recursive improvement and agent collectives: the report sets out possibilities and obstacles.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/reading/sai-review/https://smartagi.ai/en/reading/sai-review/Goldfeder and co-authors challenge definitions of AGI and propose adaptable specialisation with superhuman capabilities.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/ideas/we-are-the-neanderthals/https://smartagi.ai/en/ideas/we-are-the-neanderthals/A superior intelligence does not choose for us what deserves to be preserved.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/understand/agi/https://smartagi.ai/en/understand/agi/Tasks, comparisons, autonomy and evidence: a guide to examining AGI and superintelligence claims.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/news/chile-data-law-delay-proposal/https://smartagi.ai/en/news/chile-data-law-delay-proposal/The proposal to move the date to 1 December 2027 remains pending. Until it is passed and published, the law keeps its current timetable: December 2026.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/regulation/a-human-signs-who-decides/https://smartagi.ai/en/regulation/a-human-signs-who-decides/Mandatory review can become a ceremony when disagreement requires time, information and permissions that nobody supplies. Control over AI is determined well before the signature.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/regulation/ai-harm-insurance-answers/https://smartagi.ai/en/regulation/ai-harm-insurance-answers/Compensation can cover a loss and a certificate can establish what was assessed. Someone must still explain the decision, correct its effects and prevent the failure from recurring.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/regulation/regulating-ai-divides-market/https://smartagi.ai/en/regulation/regulating-ai-divides-market/Safety requirements cost money. What matters is who pays, what protection the spending buys and when a safeguard becomes a barrier to competition.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/regulation/algorithm-state-owes-reason/https://smartagi.ai/en/regulation/algorithm-state-owes-reason/A public decision can be fast, accurate and well explained without being justified. Chile, Europe and Canada offer different ways to prevent a score from closing the argument.Fri, 25 Sep 2026 03:00:00 +0000https://smartagi.ai/en/the-debate/agi-or-superintelligence/https://smartagi.ai/en/the-debate/agi-or-superintelligence/A machine beats us at a test. What does that establish? The debate over capabilities, promises and who sets the goals.Fri, 25 Sep 2026 03:00:00 +0000