meta launches openai while anthropic expands its platform

Key Takeaways

Sullivan & Cromwell Apologizes for AI Errors in Court Filing

Sullivan & Cromwell, a top Wall Street law firm, apologized to a New York federal judge for mistakes in a bankruptcy case document. The errors included inaccurate citations and misquoted bankruptcy code sections, which were discovered by another firm, Boies Schiller Flexner. Andrew Dietderich, a senior partner at the firm, wrote a letter admitting the mistakes and thanked the other firm for finding them. The firm stated that its AI policies were not followed and that a secondary review process failed to catch the issues. A corrected version of the filing was submitted to the court.

Sullivan & Cromwell Apologizes for AI Errors in Court Filing

Sullivan & Cromwell told a New York federal judge it made a string of errors in a document for the Prince Group case. Andrew Dietderich, co-head of the firm's global restructuring group, apologized in a letter on Saturday for mistakes including inaccurate citations. The errors were uncovered by Boies Schiller Flexner, which was also working on the case. The errors included misquoting the US bankruptcy code and citing cases incorrectly in a filing made on April 9. In multiple instances, the firm filed inaccurately summarized conclusions made in other cases using AI. The firm said it maintains policies governing the use of AI tools in legal work but admitted those policies were not followed.

Texas Must Reject Big Tech Power Grab and Regulate AI

Texas lawmakers argue that the state must reject federal attempts to block state-level AI regulations. Big Tech companies like Meta and OpenAI are pushing for federal preemption, which would prevent states from creating their own AI laws. A recent poll shows that many Americans worry AI will put people out of work. Social media platforms have already been linked to addiction among teenagers. Texas leaders believe states should have the right to create safeguards for children and families. They argue that federal preemption undermines democracy and protects the interests of Silicon Valley elites over everyday Texans.

Texas Must Reject Big Tech Power Grab and Regulate AI

Texas is at a critical juncture in its approach to regulating artificial intelligence and the power of Big Tech companies. Recent attempts by federal authorities to preempt state-level regulations threaten to undermine protections that Texas is trying to implement. The federal government has long been a proponent of a uniform regulatory framework for emerging technologies. This approach often overlooks the unique needs and concerns of individual states. One of the most pressing issues is the protection of children and families from AI algorithms used by Big Tech companies. Texas could implement stricter data privacy laws for minors and mandate transparency in AI algorithms.

Qlik Connect Calls for AI Governance and Data Quality

At Qlik Connect 2026, leaders discussed why data governance and quality are necessary steps in AI adoption. A recent report found that while 97% of enterprises have funding for agentic AI, only 18% have been fully deployed. The biggest bottlenecks on AI adoption are data quality, availability, and governance. Mike Capone, CEO of Qlik, said companies often have to fail before they understand they need a trusted data platform. He noted that CFOs are now asking where money is going and how value is being created. Qlik uses an AI governance framework to evaluate whether data is reliable enough to go into models. The company also introduced Qlik Data Catalyst to help customers build a reliable data foundation.

Anthropic and Pentagon Standoff Highlights AI Company Design

A conflict between Anthropic and the Pentagon reveals how the organizational design of AI companies affects their balance of profit and safety. OpenAI recently restructured to allow its for-profit arm to compete more aggressively in pursuit of profit. The nonprofit foundation that controls OpenAI is now legally structured like European enterprise foundations. Anthropic operates differently, with a mission-centered structure that prioritizes safety. The recent clash with the Pentagon suggests that Anthropic's governance structure is reserved for extreme circumstances. OpenAI's structure creates different incentives by allowing the company to generate and distribute surplus. This shift aligns investor interests with the nonprofit owner and gives the operating company greater strategic flexibility.

OpenAI Briefs Feds and Five Eyes on New Cyber Product

OpenAI has been briefing federal agencies, state governments, and Five Eyes allies on the capabilities of its new cyber product. Government applicants are going through the same vetting process as commercial customers for the Trusted Access for Cyber program. OpenAI is pursuing a dual-track approach by making one version of its model more widely available with strong safeguards. Another more cyber-permissive version is being released to defenders through the Trusted Access program. OpenAI Chief Global Affairs Officer Chris Lehane said this approach would allow more companies to access advanced AI tools. The company is also working with state governments to get them access to GPT-5.4-Cyber. Anthropic is currently being tested by the Pentagon, which labeled the company a supply chain risk.

Snowflake Launches Agent Framework for Finance

Snowflake's Ecosystem Agent Framework aims to automate financial services by enabling AI agents to execute tasks directly on unified data. Financial services have long optimized for data access, but the true objective has always been intelligent action. The goal now is to augment human judgment by stripping away undifferentiated work like data gathering and preparation. Historically, financial data integration was a manual process involving SFTP transfers and complex ingestion pipelines. The cloud era revolutionized this by enabling direct, live data sharing and eliminating ETL. This convergence standardized critical financial workflows involving proprietary internal context and external market signals. The industry is now entering the Agentic Era, moving from AI-ready data to Agent-Active data.

Google CEO Says 75% of Code Is Now AI Generated

Google CEO Sundar Pichai disclosed that the percentage of code written by AI has gone from 25% to 75% in exactly a year. This figure applies to complex production code across Search, Ads, Cloud, YouTube, and Android. Pichai has framed AI coding not as a replacement for engineers but as a productivity multiplier. The company writes some of the most complex production code in the world, and engineers are signing off on the AI-generated work. Co-founder Sergey Brin and Microsoft's Satya Nadella have also noted this cultural shift. The one number worth watching alongside the 75% is the approval rate of AI-generated code. Google has not disclosed how much AI-generated code gets rejected or rewritten before it ships.

Microsoft Accelerates AI Defense Against Cyber Threats

Microsoft is accelerating its work to use AI models for enterprise-grade security solutions. Recent advances in AI model capabilities are changing how vulnerabilities are discovered and exploited. AI models can autonomously discover weaknesses and produce working proof-of-concept code. These changes require organizations to rethink exposure, response, and risk. Microsoft plans to incorporate advanced AI models directly into its AI-assisted vulnerability discovery process. They will deploy detections to AI-ready posture to reduce exposure and patching. Microsoft is focusing on three areas to help customers reduce risk and improve resilience. These include AI-led vulnerability discovery, mitigations, and posture management.

Sunak Proposes Eliminating Jobs Tax to Compete with AI

Former Prime Minister Rishi Sunak suggested eliminating National Insurance to make hiring workers more attractive. He told the BBC that AI is flattening the jobs market for young people. Sunak, who is now an adviser to Anthropic and Microsoft, said concerns from graduates looking for entry-level jobs are justified. He proposed rebalancing the tax system by abolishing National Insurance over time and replacing it with taxes on corporate profits. He said these taxes would be boosted by productivity and efficiencies in deploying AI. Sunak noted that it is becoming tougher for young people to get jobs in service sectors such as law, accountancy, and the creative industries. He believes governments must tip the balance in favor of AI being used in a positive way to help people do their jobs better.

Google Adds AI Safety Features to Ads Advisor

Google is adding three new agentic safety features to Ads Advisor, its AI assistant inside Google Ads. The update introduces proactive troubleshooting, always-on security monitoring, and instant certifications powered by AI and Gemini capabilities. Ads Advisor can now flag and help resolve policy violations automatically, even before advertisers notice them. It monitors accounts 24/7, surfacing risks like suspicious domains or inactive users through a new security dashboard. Certifications that once took weeks can now be granted instantly or submitted with a single click. Instead of waiting for user prompts, Ads Advisor scans accounts and websites proactively and suggests fixes. These features are rolling out in the coming months to English-language accounts.

Sources

• An AI Screw-Up By... Sullivan & Cromwell?
• AI hallucinations found in high-profile Wall Street law firm filing
• Texas must reject Big Tech’s power grab and regulate AI
• Texas must reject Big Tech’s power grab and regulate AI
• Qlik Connect sounds call for AI governance
• Anthropic, Pentagon Standoff Shows Why AI Company Design Matters
• Exclusive: OpenAI briefs feds and Five Eyes on new cyber product
• Snowflake's Agent Framework for Finance
• 75% Of Code At Google Is Now Generated By AI: CEO Sundar Pichai
• AI-powered defense for an AI-accelerated threat landscape
• Eliminate jobs tax for workers to compete with AI, says Sunak
• Google rolls out new AI safety features in Ads Advisor