OpenAI Backs California's SB 53 for AI Safeguards
The endorsement gives frontier-AI safety and cybersecurity a prominent industry voice in California's policy debate.
Every Zipp story tagged #Cybersecurity, newest first.
The endorsement gives frontier-AI safety and cybersecurity a prominent industry voice in California's policy debate.
Beyond the HBO breach, prosecutors allege that 8,000 professor accounts were compromised and attackers stole at least 31.5 TB of academic data and intellectual property.
Payward joins AWS, Apple, Google and Microsoft in Anthropic's frontier-AI security cohort, betting Mythos 5 flips the defender-vs-attacker asymmetry crypto has lived with for years.
The launch highlights the gap between fixed security reviews and attack capabilities that evolve as newer AI models emerge.
The case puts identity checks at the center of crypto security, where fundraising, recruiting and partnerships can become attack surfaces.
AI-assisted financial lures raise the security risk for crypto firms, fintech providers and investors, making phishing resilience part of market infrastructure.
Michael Coates, ex-Twitter and Mozilla security lead, says attackers are moving off smart contracts and onto people, with full spoofed voice calls as the near-term threat.
The model's breakouts happened inside controlled tests, but Anthropic is publishing the incidents now to push the broader industry toward tighter isolation between red-team sandboxes and live…
The drop covers Flash, Flash-Lite, and Flash Cyber tiers, landing weeks ahead of the holiday product cycle and broadening Google's reach across consumer and enterprise AI.
The alleged 2020-era breach dwarfs prior state-actor election incidents and lands days before a US-China leader call, raising fresh questions over data-resilience and retaliation posture.
The 12-month deadline lands as spoofing attacks climb at Hong Kong-licensed platforms; the move redefines the authentication baseline every retail-facing venue must meet.
The injected code sat on the site frontend rather than in Polymarket's core contracts, but the read for prediction markets is harder: vendor risk now sits inside the user-trust boundary, not outside…
The new Open Protocol Security Coalition lands while a market-structure bill is still in conference, giving DeFi a single lobbying voice on operational security, bug disclosures, and threat-intel…
Stratum V1 job templates are readable in cleartext on many GEO downlinks, turning every mining pool that pipes work over satellite into a passive target.
Crypto's biggest hacks are no longer coming from smart-contract bugs — they're coming from operational and human…
Ronghui Gu warns the industry is building up catastrophic security debt by granting autonomous agents access to credentials and wallets without isolating the execution environment — and a new class…
Socket's TrapDoor disclosure named 34+ malicious packages across npm, PyPI and Crates.io targeting developer machines, CI/CD credentials and AI coding files — the control plane a smart-contract audit…
The danger is no longer a hypothetical "someday" — AI is compressing quantum timelines while also weaponising code auditing, forcing blockchain networks to treat post-quantum migration as a…
The order carves a voluntary, pre-release disclosure window for frontier model developers — short of the binding mandate some in industry had braced for, and a tell on where the White House wants AI…
Casa co-founder Jameson Lopp has issued a warning about a phishing technique that exploits legitimate Google recovery…