Okta Launches Blueprint Alliance for Agentic AI Security

Key Takeaways

• Okta launched the Blueprint Alliance to secure agentic AI through shared industry safeguards, giving each agent its own identity and monitored permissions.
• Oracle is updating data security controls in its latest database version, emphasizing identity protection as AI agents carry permissions on behalf of individuals.
• Cloudflare reports over 50 percent of its Q2 network traffic was nonhuman, and more than 80 percent of major AI companies are already customers.
• Bill Gates said global AI regulation is harder than nuclear arms talks and urged the U.S. to lead international consensus efforts.
• The UK Games Expo banned AI-made games, art, books, and craft items from its exhibition floor after backlash over promoting an AI-built game.
• Europe faces regulatory confusion combining the EU AI Act's risk-based approach with aviation safety rules for AI-powered drones.
• AI still struggles to predict hurricane intensity, with rapid intensification events remaining difficult due to chaotic atmospheric dynamics and training data uncertainties.
• Palmer Luckey says AI does not make founding a company easier, as legal issues, big decisions, and hiring remain core challenges.
• Security experts warn Q-Day, when quantum computers break encryption, could arrive within five years, urging immediate prioritization of critical data protection.
• John Binks' new book AI Made Simple, available on Amazon, explains AI concepts including machine learning, LLMs, and AI agents in plain English.

Okta launches Blueprint Alliance for agentic AI security

Okta introduced the Blueprint Alliance to help secure agentic AI across multiple platforms. The approach calls for shared industry safeguards instead of isolated vendor controls. Each AI agent would get its own identity, limited permissions, and monitored behavior. This lets organizations spot when an agent drifts from its intended task. Okta's Wylie stressed that the industry must work together to protect the full agentic stack.

Oracle tackles agentic AI security through identity and data controls

Oracle highlighted new pressures on identity and database security as agentic AI spreads in security operations. Konstantas explained that agents carrying permissions on behalf of individuals make identity protection more important than ever. Oracle is updating its data security capabilities in its latest database version. The company urges customers to first clarify their goals and locate sensitive data before applying safeguards. Strong, uniform identity controls remain key to managing agent access safely.

Debate grows over regulating artificial intelligence risks

Commentary across major outlets debates how to handle AI's growing risks. Some argue that Republicans and Democrats must unite for a shared approach to AI safety. Others say lawsuits and legal liability already provide enough deterrent against reckless AI products. Ezra Klein suggests banning AI models from writing their own code to prevent runaway self-improvement. The discussions reveal deep disagreement on whether regulation or legal consequences best address AI dangers.

AI struggles to predict hurricane intensity changes

AI has advanced weather forecasting overall but still struggles with predicting hurricane intensity. Rapid intensification events like Hurricane Matthew in 2016 show how hard these storms are to forecast. Regional predictions are especially difficult because AI must capture fast-developing extreme events. Simulated data used to train models contains approximations and uncertainties. Chaos theory also plays a role, as tiny changes in a hurricane's initial state can cause big differences in predicted intensity.

Palmer Luckey says AI won't simplify founding a company

Palmer Luckey stated that AI will not make a founder's job easier. He pointed out that difficult challenges like handling legal issues, making big decisions, and hiring the right people remain unchanged. The message stresses that AI tools do not remove the core demands of running a business. Founders still face complex responsibilities that technology alone cannot solve.

UK Games Expo bans AI-made games and art from convention floor

The UK Games Expo banned products mostly made using AI from its exhibition floor. The rule covers art, text, games, books, craft items, and clothing. The only exception is basic spell checking and minor editing. The policy came after the expo faced backlash for promoting an AI-built game and hiding negative comments online.

New book AI Made Simple explains artificial intelligence in plain English

Author John Binks released a book called AI Made Simple to help everyday readers understand AI. The book covers topics like machine learning, generative AI, large language models, and AI agents without using heavy technical jargon. It also discusses privacy, security, and responsible AI use. The book is available through Amazon and other retailers worldwide.

Europe struggles to combine AI rules with drone regulations

Europe has two major rule systems that were not designed to work together: the EU AI Act and the European drone framework. The AI Act uses a risk-based approach while aviation rules focus on safety. This creates confusion for manufacturers building AI-powered drones. EASA is working on new proposals to address AI trustworthiness in aviation, starting with human assistance and human-AI cooperation levels.

Bill Gates says global AI rules are harder than nuclear arms talks

Microsoft co-founder Bill Gates said agreeing on global AI regulations will be more difficult than Cold War-era nuclear arms talks. He spoke after President Donald Trump made related comments. Gates believes the U.S. should lead efforts to reach international consensus on AI. He pointed to past global successes like aviation safety and ozone layer protection as examples.

AI security demand could boost Cloudflare's Zero Trust and SASE growth

Cloudflare is seeing more interest in its Zero Trust and SASE products as companies try to secure AI agents. Management said large enterprises are reaching out because they want to adopt AI safely. The company is building its security model around AI agents rather than just human users. Over 50 percent of Cloudflare's network traffic in the second quarter was nonhuman, and more than 80 percent of major AI companies are already Cloudflare customers.

Security Experts Warn to Prioritize Quantum Prep Amid AI Threats

Security teams must prepare for quantum computer threats even while dealing with new AI security challenges. Experts say waiting for a specific date when quantum computers break encryption could leave organizations with too little time to react. Some estimates suggest this event, known as Q-Day, could happen within five years. Industry leaders like Chris Konrad of World Wide Technology report that customers are now urgently planning for long-term cryptographic resilience. Experts advise companies to ruthlessly prioritize their most critical data and start the transition process immediately rather than waiting for a perfect plan.

Sources

• Okta advances agentic AI security through Blueprint Alliance
• Agentic AI in security operations: Oracle on data and identity
• Could regulations reduce the risk from Artificial Intelligence?
• Why AI has trouble predicting the fury of hurricane intensity
• Palmer Luckey says AI won't make a founder's job easy
• UK Games Expo bans games and art made mostly using AI — UKGE enforces booth shutdowns and bans after deleted-comment debacle, says that the ‘central creative process must be conducted by humans’
• New Book 'AI Made Simple' Makes Artificial Intelligence Easier to Understand for Everyone
• When AI meets U-space – the regulatory collision Europe has not yet resolved
• Bill Gates says global AI rules will be harder than nuclear talks
• Can AI Security Demand Boost Cloudflare's Zero Trust & SASE Growth?
• AI May Be Dominating Cybersecurity, But Quantum Preparations Can’t Wait: Analysis