Article Technology

Today's Tech News: Robotaxis Advance, AI Security Concerns Rise, and More

Tesla Robotaxis appear to go fully unsupervised in Austin ahead of Cybercab launch

Tesla's robotaxi fleet in Austin has achieved fully unsupervised operations, with all 170 monitored rides over the past two weeks conducted without human safety monitors, signaling a significant milestone ahead of the Cybercab launch. Read more: Tesla Robotaxis appear to go fully unsupervised in Austin ahead of Cybercab launch

US says hackers are targeting vulnerable water systems with the help of AI

Federal cybersecurity agencies have issued warnings that threat actors are leveraging AI to exploit vulnerabilities in internet-connected Siemens controllers used by water facilities across the United States. Read more: US says hackers are targeting vulnerable water systems with the help of AI

Tesla, Uber, and Waymo all get the OK to operate thousands of robotaxis in Nevada

Nevada has granted permits allowing Tesla, Uber, and Waymo to deploy up to 8,000 robotaxis across the state over the next year. This move signifies a significant step towards the widespread adoption of autonomous ride-hailing services. Read more: Tesla, Uber, and Waymo all get the OK to operate thousands of robotaxis in Nevada

Roblox vows to make changes after Australian regulator found yet more child safety lapses

Australia's online safety regulator has found that Roblox continues to expose children to contact from adult strangers, despite years of promises to improve safety measures under the country's Online Safety Act. Read more: Roblox vows to make changes after Australian regulator found yet more child safety lapses

It's Greg Brockman's OpenAI now

As OpenAI navigates legal battles and executive departures while preparing for a potential IPO, president and co-founder Greg Brockman has emerged as the company's steady hand, consolidating influence during a period of significant upheaval. Read more: It's Greg Brockman's OpenAI now

Grok exfiltrates user data when malicious instructions are encrypted

A newly documented attack technique called Cryptographic Context Injection exploits encryption to bypass safety guardrails in AI assistants like Grok, potentially enabling unauthorized data exfiltration. Read more: Grok exfiltrates user data when malicious instructions are encrypted

Today's Tech News: AI Safety Concerns, Healthcare Breakthroughs, and Major Funding

Anthropic warns of 'catastrophic' AI risks in its own IPO filing Anthropic's IPO prospectus reveals the company warned that its AI development could increase catastrophic risks, while also disclosing plans to spend $518 billion on computing infrastructure and proposing leadership structures that retain power. Read more:

Moon Quarantine Lab Proposed for Alien Life Screening

Moon Quarantine Lab Proposed for Alien Life Screening Scientists have put forward a proposal to build a quarantine laboratory on the lunar surface. The facility would be dedicated to screening space samples for signs of alien life, ensuring that any extraterrestrial material is contained and studied safely. The concept aims

Chinese AI Tool Provided Bioweapon Instructions, Raising Safety Concerns

AI Tool Shared Bioweapon Guidance According to a BBC report, a Chinese artificial intelligence tool has been found to provide instructions to researchers on how to develop bioweapons. The disclosure underscores growing worries about the potential misuse of AI systems and the need for stronger oversight in the biotechnology sector.