OpenAI Introduces Lockdown Mode to Combat Prompt Injection Attacks
OpenAI announced Lockdown Mode, a new feature designed to protect sensitive data from prompt injection attacks, effective June 6, 2026.
ARTICLES BY NADIA RAHMAN
LATEST NEWS
-
-
Sriram Krishnan Steps Down From White House AI Role
Former tech executive Sriram Krishnan is leaving his White House AI advisor role at the end of June, citing his work on the administration’s AI Action Plan.
-
Chinese Crypto-Funded Peptide Labs Are Booming
Crypto-funded Chinese labs are now selling peptides, a $100 million industry, after pivoting from fentanyl production, according to Chainalysis.
-
Florida Lawsuit Labels ChatGPT as Defective Product and Public Nuisance
Florida sues OpenAI and Sam Altman, alleging ChatGPT endangers minors and promotes harmful behavior.
-
Microsoft CEO Criticizes Plan to Make AI Agent Addictive
Microsoft CEO Satya Nadella rebuked an internal proposal to design AI agent Scout as deliberately addictive, calling it 'absolutely not a goal.'
-
Elon Musk seeks to end FTC data-audit order on X
Elon Musk is trying again to escape a 20-year data-privacy order from the FTC, which he claims is outdated and costly.
-
Meta Hack Highlights AI Security Vulnerabilities
Attackers exploited Meta's AI customer support agent to steal Instagram accounts, including the dormant Obama White House account, by tricking it into changing email addresses.
-
AI Tools Expanding Access to Justice in Courts
A study shows the share of lawsuits by self-represented individuals rose from 11% in 2022 to 16.8% in 2025, with AI use playing a key role in this trend.
-
AI Systems Now Coach Amateur Virologists, Tech Leaders Push for DNA Security Laws
Top tech leaders and scientists urge Congress to mandate synthetic DNA order screening, citing AI's ability to outperform virologists on lab procedures.
-
UK Orders Google to Improve AI Search Links and Allow Publishers to Opt Out
UK regulators have ordered Google to enhance attribution and allow publishers to opt out of AI features in search. The CMA requires clearer links and opt-out mechanisms within nine months.
-
Trump's AI executive order may not prevent dangerous deployments
Trump signed an executive order that promises voluntary safety testing for frontier AI models but critics say it lacks enforceable measures.
-
xAI Seeks to Reveal Deepfake Victims' Names in Lawsuit
xAI, Elon Musk's AI firm, wants to strip pseudonyms from four deepfake victims in a federal class-action lawsuit, citing public interest in the case.