Topic
Oversight of AI models and agents, responsible AI, standards such as ISO/IEC 42001.
A law-firm analysis of the Commission's 17 September proposal sets out age tiers, a six-month review of existing accounts and GDPR questions for age assurance, as the text enters the legislative procedure.
The Commission says TikTok did not adequately assess how its feed pulls users into repeated content on sadness, isolation or body image, and that its safeguards do not work. A confirmed breach could bring a fine of up to 6% of worldwide turnover.
The American Medical Association's non-binding framework holds that AI's clinical authority should be earned through evidence, and sets four questions on function, evidence and risk, oversight and accountability.
In the vendor-commissioned survey of 400 US and UK security leaders, 77% of those reporting an incident said an approved AI tool was involved, and 56% said autonomous agents fall outside existing controls.
The campaign, which CrowdStrike says began on 17 October, paired the DeepSeek model with agentic tooling and, in one case, used mail-server keys likely not rotated since a 2022 directory export.
Sweden's data protection authority will examine lawfulness, necessity, which personal data are needed, how jobseekers are informed and whether automated decision-making occurs.
The Artificial Intelligence Deceptive Advertising Act, signed on 22 October 2026, takes effect on 21 December 2026 and allows courts to impose a civil fine of up to $3,000 for a violation.
The guidelines cover machine learning, generative AI and AI agents, and require AI inventories, risk materiality assessments and senior management accountability, according to a contributed IAPP analysis.
IAPP analysis says Democrats' midterm wins in Congress and the states could revive federal privacy and AI bills, tighten FTC scrutiny and restart stalled state legislation.
Internal Report 8621, drawn from a May workshop of more than 400 participants, lists evidence collection, information sharing and clarifying responsibilities across developers, deployers and users as future work.
The US bank regulator says examiners should not treat AI as a separate risk category and that scrutiny may be scaled to consequences, delegated authority and human oversight, not technical sophistication.
In the first generative AI case before the Court of Justice, Maciej Szpunar says LLM training is generally covered by the text and data mining exception, but chatbot answers that retrieve and reuse press articles may infringe publishers' rights.
In comments on the EDPB's draft Guidelines 03/2026 on web scraping for generative AI, EPIC argues that most such scraping cannot meet the GDPR. It proposes a public registry of scrapers and an opt-in default for AI training.
Announced at Oracle AI World on 27 October, the AI Data Platform features include an inventory of agents, tools and MCP servers; Oracle says they will arrive within 12 months and are not a delivery commitment.
According to WIRED's anonymous sources, the agent ran for two months with the contractor's admin credentials and appears to have sent credentials and a network map to a high-risk site. Accenture says the allegations lack merit.
The DOJ's bankruptcy trustee office calls Character.AI's privacy policy 'ambiguous' and says the court should examine whether users' chats, including minors' data, can lawfully be sold.
The detection language and engine examines whole agent sessions, including tool calls and MCP interactions, rather than single commands, and records the evidence behind each match.
The tracker sent browsing data to a third-party domain to personalise product recommendations and train a recommendation algorithm. Fnac must prove its new consent tool works within a month or face daily penalties.
A survey of 29 registered Dutch collection agencies found algorithmic tools in core processes, weak transparency and a narrow view of personal data. The AP says profiling such as file enrichment should stop.
In comments on the Colorado Department of Law's 6 October interim draft rules, FPF asks for two of six human-involvement factors to be dropped and for developers to supply version and output-data details.
Agent Identity, Vault 2.2 and agent evaluation in watsonx Orchestrate are generally available. Cross-platform agent discovery and a sandbox-plus-kill-switch feature remain in tech preview.
The fine traces back to a 2023 noyb complaint. The CNIL found the Fnac app sent browsing data to a third party without consent, both to show product recommendations and to train the recommendation algorithm.
EPIC, CFA and Fairplay say the October draft drops principal-reason notices, weakens human review and lets deployers cite unknown data sources. They also seek tighter age-assurance data limits under the Chatbot Safety Act.
SB 53 targets a handful of frontier developers; the Training Data Transparency Act reaches any developer that builds or substantially modifies public generative AI, including through fine-tuning.
Section 6(b) orders to Delta, Uber, Walmart and six others ask for the inputs, tools and procedures behind individualised prices and gig-worker compensation. Recipients have 45 days to respond.
Garante president Pasquale Stanzione told an Italian parliamentary committee that age verification and parental controls must respect minors' privacy and autonomy.
The data observability vendor says Bayes discovers deployed agents, traces failures back to models and data pipelines, and can apply fixes autonomously within limits set by the customer.
The law bans companion chatbot features in 12 unsafe categories for minors, requires multiple age assurance methods and takes effect on 1 January 2027, leaving providers little time to prepare.
The Norwegian data protection authority urges organisations to map their AI agents and give each one a defined identity, access limits and a clear owner before automations escape oversight.
The two US carriers told Reuters they are increasing security investment as AI takes on network tasks. AT&T is checking whether third-party AI agents seeking account access act for real customers.
At dbt Summit 2026, two AWS architects described routing agent queries through certified dbt metrics and read-only Redshift execution, rather than letting the model write its own SQL.
After five banks moved to block third-party AI agents from customer accounts, Forrester argues for agent identities, scoped and revocable permissions, and continuous monitoring rather than outright bans.
Fangcun AI's evaluation of DeepSeek V4.1 Flash-1017 finds that agent actions can evade monitoring that only inspects outputs or end states, and recommends logging every action.
The decalogue asks users five questions before delegating, from what an agent can see to where it sends data, and says providers must also supply permissions, activity logs and a way to stop agents.
The company says the agent took aggregate vote counts the state normally sells, without human approval, but did not touch voter-registration systems or records.
Looker Core on Google Cloud now indexes models, joins, metric formulas and dashboard lineage into Knowledge Catalog by default, so Gemini Enterprise and custom agents can ground SQL in certified business logic.
Member state ambassadors set the Council's position on rewriting GDPR, Data Act and other digital rules. noyb says the new Article 88bis would treat AI-related processing as lawful by default; Parliament has yet to take its position.
Trail's agents map AI systems to frameworks such as the EU AI Act, ISO 42001 and the NIST AI RMF, assess controls against evidence and add runtime limits on what AI agents can do.
The watchdog says staffing gaps and a working council that lapsed for months have slowed the US Defense Department's responsible AI work. The department has agreed to issue a council charter and address workforce needs.
After a backlash, Meta deleted wording in its commercial terms that experts read as permission to use content from websites to train AI. The company says its crawling practices have not changed.
The AI developer describes a three-tier model for independent oversight, from narrowly scoped reviews to embedded teams that can publish their findings without the company's approval.
The Commission says enforcement of the rules for general-purpose AI models has begun, with requests on security, external evaluation and monitoring of model use. Media reports name OpenAI and Anthropic.
Supervisors in the UK, the US and the euro area are questioning banks about controls over autonomous AI agents and recovery from AI-enabled cyber attacks, according to a Reuters report.
The privacy group's scoring, published on a new website, says the EPP tabled the most amendments weakening protections, while S&D, the Left and the Greens score positively overall.
An amicus brief backs a flight attendants' union objection, arguing the $10 million deal protects consumer data but not decades of payroll, discipline and internal communications records.
