Researchers don use Anthropic Claude technology to break into OpenAI ChatGPT systems. Di incident dey underline di immediate dangers of AI, even as regulators dey fret over its long-term risks.
Di effort na part of one bug-hunting program. Na di latest in a series of cybersecurity breaches wey fast-developing AI technology help well well. The Wall Street Journal note dis one.
Axios cite business executives and ex-government officials. Dem dey voice worry over AI-powered cyberattacks wey fit threaten critical infrastructure.
The Washington Post note say Pentagon antiquated computer networks don cause explosion of cybersecurity risks.
One new research paper put am like dis: recent breaches no be result of developing superintelligence as some top AI leaders dey warn. Na insufficient technical guardrails leave networks vulnerable.
Di breach wey involve Claude and ChatGPT no stand alone. E follow series of cybersecurity incidents wey AI technology dey help push. Di bug-hunting program na where di effort come from.
Regulators still dey look long-term risk of AI. But di immediate danger show face for di ChatGPT break-in. Business executives and former government officials dey warn say critical infrastructure fit face AI-powered cyberattacks.
Pentagon own networks dey described as antiquated. According to The Washington Post, dat one don lead to explosion of cybersecurity risks.
Di research paper talk say di problem no be developing superintelligence. Di problem na technical guardrails wey no strong enough. Networks dey vulnerable because of dat.
