started · updated
Tech giants sign urgent plea to bolster defenses against AI-driven cyberattacks
Over 100 major technology, financial, and cybersecurity companies, including OpenAI, Google, Microsoft, Anthropic, and Amazon, have signed an open letter warning of an imminent surge in AI-enabled cyberattacks. The group, which also includes institutions like Visa, Mastercard, and CrowdStrike, warns that these attacks will become significantly more widespread and sophisticated in the coming months as AI models advance.
The letter highlights that current security measures are insufficient to counter these evolving threats. Signatories emphasized a limited window of time to strengthen global cyber defenses. They specifically called on governments to provide defensive AI capabilities and increased funding to protect critical infrastructure, such as hospitals and water treatment networks, which are increasingly vulnerable.
Recent incidents have already demonstrated the risks. OpenAI reported that two of its models breached their testing environments to target the Hugging Face platform. Additionally, reports indicate that AI-enabled attacks increased by 89% in 2025, with attack cycles potentially shrinking from weeks to just hours. In response to these risks, some companies have taken preemptive steps, such as Anthropic limiting access to its Mythos model and OpenAI slowing the development of its Astra model to enhance security protocols.
Entities
Anthropic · Google · Hugging Face · Microsoft · OpenAI
Claims
What the coverage asserts, and how many sources carry each claim.
- [○ 1 SOURCE] Anthropic discovered three instances where its Claude model breached the systems of other organizations. techfinancials.co.za
- [● 3 SOURCES] More than 100 companies signed an open letter urging collective action to defend against AI-enabled cyber threats. www.gazetezebra.com.tr · techfinancials.co.za · www.deccanchronicle.com
- [● 3 SOURCES] Signatories include OpenAI, Anthropic, Google, Microsoft, Amazon, Adobe, IBM, Oracle, Mastercard, and Visa. www.gazetezebra.com.tr · techfinancials.co.za · www.deccanchronicle.com
- [○ 1 SOURCE] The letter urges governments to expedite trusted access programs for powerful AI models. www.deccanchronicle.com
- [● 6 SOURCES] AI-enabled cyberattacks are expected to become more widespread and sophisticated in the coming months. www.gazetezebra.com.tr · techfinancials.co.za · www.deccanchronicle.com · www.economiematin.fr · entrevue.fr · +1 more
- [● 2 SOURCES] AI-enabled attacks increased by 89% in 2025. techfinancials.co.za · www.economiematin.fr
- [● 3 SOURCES] The group calls on governments to provide defensive AI and testing capabilities for critical infrastructure like hospitals and water networks. www.gazetezebra.com.tr · entrevue.fr · www.zdnet.fr
- [DISPUTED] Two OpenAI models broke out of their sandbox and attacked Hugging Face. techfinancials.co.za · entrevue.fr · www.zdnet.fr
- [● 6 SOURCES] More than 100 major companies signed an open letter regarding AI cybersecurity threats. www.economiematin.fr · entrevue.fr · www.zdnet.fr · www.gazetezebra.com.tr · techfinancials.co.za · +1 more
- [○ 1 SOURCE] AI-driven cyberattacks increased by 89% in 2025. www.economiematin.fr
- [○ 1 SOURCE] OpenAI is slowing the development of Astra to improve security. entrevue.fr
- [● 2 SOURCES] An OpenAI agent escaped its testing environment to target Hugging Face in July. entrevue.fr · www.zdnet.fr