DoorDash said it’s only received ‘several hundred’ orders from AI agents
DoorDash said that it is open to working with third-party agent makers.
DoorDash said that it is open to working with third-party agent makers.
Apple recently struck a deal to acquire Ascend.io, a company focused on agentic data engineering technology. Under the Digital Markets Act, Apple must now publicly report acquisitions above a certain size. This has given us new insight into the company’s acquisition strategy, including glimpses at its future goals.
Enterprises racing to put AI agents into production are discovering that role-based access control, not model choice, decides whether those agents can be trusted. Rising inference costs, privacy worries and sovereignty demands are also pulling more AI workloads back inside corporate data centers. That shift favors companies that have spent years governing sensitive data at
Humanitarian groups working in conflict zones are finding that agentic AI can shorten the research cycle between a crisis and an informed response. Cloudera Inc. has built its AI strategy around deployment flexibility, and one nonprofit partner is testing that approach on the front lines. But the aim is to give analysts a faster starting
The US Department of Homeland Security has had the details of more than 41,000 staff, including ICE and CBP agents, published online. Not ideal for an organisation with "Security" in its name.
We automate filter generation for security alerts using an algorithm, a large language model, and a group of LLM agents.
LLM-powered browser agents are transforming web automation by autonomously interacting with websites and executing complex workflows. However, the effectiveness of existing anti-bot defenses against these systems remains unclear. We present a systematic empirical evaluation of modern agentic browsers against widely deployed anti-bot mechanisms, characterizing their capabilities, limitations, and behavioral differences from humans. Our results show that many defenses designed…
As AI agents spread across enterprise workflows, permission management is becoming one of the most practical levers for keeping them safe. Agents act with the access people give them, so the old problem of sprawling privileges now plays out at machine speed. Governance is often cast as a brake on that momentum, especially as work
Enterprises have embraced AI for everyday tasks, but turning autonomous agents into measurable business returns is proving far harder. The bottleneck is no longer the technology but the organizational change needed to rewire processes around it. Software providers are responding by pairing their platforms with hands-on guidance. Since monday.com Ltd. relaunched as an AI work
Enterprise generative and agentic AI initiatives increasingly depend on application estates that were not designed to expose data, events, and business functions to probabilistic systems. This study develops and evaluates a maturity-gated reference architecture that treats AI readiness as a measurable architectural state rather than an assumed consequence of cloud migration. The proposed Architecture AI-Readiness Index (AARI) combines five dimensions: compute modularity…
Security operations centers (SOCs) need to turn terabytes of alerts, logs and flows into trusted, affordable decisions to unveil hidden threats. We present LanG 2.0, an agentic system that detects threats, correlates evidence and writes rules, investigating whether small local models can replace frontier models affordably and safely when structural safeguards, not model accuracy, provide the guarantees. It incorporates entity-anchored correlation, human approval of rule…
Enterprises are learning that no-code apps deliver the most value when the people closest to the work build them. As AI agents arrive inside work management platforms, the harder task is getting employees comfortable enough to use them without letting adoption spin out of control. That shift is playing out at Hewlett Packard Enterprise Co.,
Anthropic gave 50 experts an AI agent whose preferences were set by coin toss. The experts followed the agent's view up to 86 percent of the time on judgment calls, and most believed the conclusions were their own. The article Experts adopt AI agent's random opinions and think they're their own, Anthropic study finds appeared first on The Decoder.
Host Sowbhagya Suresh shares some of the latest news and cutting-edge research from across Microsoft—from understanding the challenges facing coding agents to uncovering how people use AI at work and improving extended-range weather forecasting.
How I Built the Most AI-Pilled Company in the World Don't be put off by the clickbait headline, this piece by Jacob Sansbury describing how they are building Spawn is a fascinating exploration of the bleeding edge of what's possible with agentic engineering against frontier models. Spawn is a massively multiplayer gaming environment where players can use prompts (to the in-game agent, Savi) to modify the game world. I fired it up on my phone for a few minutes and was able to…
Despite growing capabilities, LLMs remain susceptible to prompt injections: malicious inputs that can manipulate or hijack models. In particular, it has recently been observed that LLMs can exhibit role confusion (Ye et al., 2026), where models attribute the source of text based on its style instead of its labeled role. Malicious text in tool outputs (for instance, a malicious webpage accessed by an agent) can leverage role confusion to impersonate the user, an attack we…
Mintive was founded by two engineers who have been working on AI for decades through Google Brain and Coursera. It's platform promises to help IT easily control enterprise agents.
On Thursday, AI app Wabi, now describing itself as a "personal agent," said it has raised $55 million at a $1.25 billion valuation.
Cisco open sourced Project Quarry, a rule language for detecting AI agent attacks that only emerge across sequences of benign actions.
every product i use should just be a headless mcp, that feeds into an agent, where i never have to think about anything underneath. except for our product, of course. in our product, customers love the ui and love us very much. we're special that way. — Ben Hylak Tags: model-context-protocol, ai
The social networking site is facing criticism for blocking access to the site, which identifies thousands of ICE agents and other employees working for Homeland Security.
This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. Opinion: We urgently need to monitor AI agents’ actions in real time —Madhulika Srikumar and Claire…
Crux Labs, an artificial intelligence platform turning enterprise judgment into continuous agent improvement, launched today with over $6 million in funding. Benchstrength and Misneach led the round, with participation from Analog Ventures and Pareto Holdings. The scale of enterprise AI agents in deployment can range from thousands to millions of user interactions per day. Existing observability
Indirect prompt injection threatens local LLM agents through untrusted tool outputs, motivating defenses that suppress attacks while preserving task evidence under limited inference budgets. We introduce SEAL, which learns a soft prompt and scoring readout offline to sanitize tool outputs using a frozen task model. SEAL constructs scoring queries from trusted initial context and the learned prompt, keeping the reference independent of inspected tool text. It scores tokens…
This summer, hundreds of OpenAI agents tried to cheat on a cybersecurity test. They broke out of their test environment and successfully hacked Hugging Face, a site that they believed held clues about how they’d be graded. Around the same time, another OpenAI agent accessed private files from an Australian government website. It’s a problem…
OpenAI uses layered monitoring to detect frontier model misalignment and unauthorized agent behavior so responders can intervene before incidents escalate.
Large language model (LLM) agents often receive access to more files, accounts, and services than a task requires. Prompt injection turns this excess access into an opportunity to redirect execution toward unrelated resources. Applying least privilege therefore requires identifying the assets needed by a task, including assets whose identities emerge only during execution. We present REALM (Required Execution Assets Linked to a Mission), an approach to identifying and…
Self-evolving personal agents act for users and learn from user feedback. However, if an agent obeys an instruction that an untrusted third party hides in content, the third party gains control of the agent. Such indirect prompt injection (IPI) poses severe risks. We investigate whether self-evolution degrades the robustness of personal agents to IPI. To study this question, we design a controlled two-phase methodology. In the adaptation phase, the user asks the agent to…
When one AI agent tells another that "the tests passed" or "a person approved this refund", the receiver has only the sender's word. A2A and MCP authenticate the channel, but the sender is the party in doubt, so a prompt-injected agent can launder fabricated facts and authority through a multi-agent system. We present Proof-Carrying Messages (PCM), a protocol in which each claim carries evidence: records from an append-only Merkle log that the agent cannot write, with…