Legacy Code

Adobe has launched an AI assistant for Photoshop on the web that allows users to edit images by describing what they want in plain language. Instead of navigating menus and tools, users can type instructions like "remove the background" or "make the sky look like sunset" and the assistant will execute the changes directly. The feature is currently in beta.

Amazon has called a mandatory engineering-wide meeting and is implementing a new policy requiring senior engineers to review and sign off on any code changes that were written or assisted by AI tools. The move comes after a series of service outages that the company has linked, at least in part, to AI-generated code being deployed without sufficient human oversight.

Amazon has launched a healthcare-focused AI assistant directly within its main website and mobile app. The tool can answer health questions, help users understand their health records, manage prescription renewals, and book medical appointments. The assistant integrates with Amazon Pharmacy and Amazon Clinic, the company's telehealth service.

Anthropic has launched a Code Review feature within its Claude Code platform, designed to address a problem that is becoming acute at many software organizations: the volume of AI-generated code is growing faster than human teams can review it. The tool uses a multi-agent approach, meaning several AI systems work in parallel to analyze code, flag logic errors, identify security issues, and surface inconsistencies before a human reviewer sees it.

Anthropic, the AI company behind the Claude family of models, has filed a lawsuit against the U.S. Department of Defense after the agency designated it a supply-chain risk. The designation — typically reserved for companies with ties to foreign adversaries — effectively places Anthropic on a restricted list, making it difficult or impossible for federal contractors and agencies to use its products.

Google has rolled out a significant update to its Workspace productivity suite, embedding its Gemini AI assistant more deeply into Docs, Sheets, Slides, and Drive. The new capabilities include a Gemini chat panel inside Google Docs, the ability to generate entire spreadsheets from a text description, and an AI-powered search feature in Drive that understands natural language queries.

Grammarly, which acquired email productivity app Superhuman, has been using the real names and professional identities of published authors and journalists to power an "Expert Review" feature — without asking for their permission first. The feature presents AI-generated editorial feedback styled as coming from a named human expert, lending the tool a credibility it derives from real people's reputations without their knowledge or consent.

Legora, an AI platform built for legal professionals, has closed a $550 million Series D funding round led by Accel, pushing its valuation to $5.55 billion. The company is using the capital to accelerate its expansion in the United States, where it sees the largest addressable market for AI-assisted legal work.

OpenAI has acquired Promptfoo, a startup that specialized in testing and securing AI systems against vulnerabilities such as prompt injection attacks, jailbreaks, and unintended outputs. The acquisition is aimed at strengthening the security posture of OpenAI's growing portfolio of AI agents — automated systems that take actions in the real world on behalf of users or businesses.

Yann LeCun, Meta's former chief AI scientist and one of the field's most prominent skeptics of large language models as a path to human-level intelligence, has raised $1.03 billion for his new startup AMI Labs. The round is reportedly the largest seed funding in European history. LeCun's core argument is that truly capable AI must understand the physical world — how objects behave, how cause and effect work in three dimensions — not just predict the next word in a text sequence.

YouTube is expanding its AI-powered deepfake detection tool — previously available only to content creators — to a pilot group of politicians, government officials, and journalists. The tool allows these individuals to identify and flag unauthorized AI-generated likenesses of themselves appearing in YouTube videos, with the platform then reviewing flagged content for removal.

Zoom has announced a major expansion of its AI capabilities, bundling a full office productivity suite into its platform alongside two features that signal where business communication is heading: AI-generated avatars that can attend meetings on your behalf, and real-time deepfake detection built directly into calls.