# AI's security crisis sparks urgent defense scramble

> Major companies are racing to patch cyber vulnerabilities as AI risks escalate, but experts debate whether slowing development is the right move.
- Title: AI's security crisis sparks urgent defense scramble · Pymnts
- Summary: Major companies are racing to patch cyber vulnerabilities as AI risks escalate, but experts debate whether slowing development is the right move. Companies…
- Keywords: citigroup, jane fraser, artificial intelligence, cybersecurity, anthropic, mythos model, ai hacking, cyber defense, ai development, wall street, business, finance
- Source: Pymnts — https://pymnts.com/cybersecurity/2026/citi-ceo-says-companies-scrambling-to-launch-ai-cyber-defenses
- Author: PYMNTS
- Published: 2026-09-20T23:51:57+00:00
- Read time: 1 min
- Topics: citigroup, jane fraser, artificial intelligence, cybersecurity, anthropic, mythos model, ai hacking, cyber defense, ai development, wall street, business, finance
## Companies racing to patch AI vulnerabilities
Citigroup CEO Jane Fraser says firms worldwide are in a "tsunami of patching" to defend against AI-related cyber threats. The urgency stems from recent AI security breaches and warnings from major tech companies.

> There is a tsunami of patching going on in the world at the moment in all companies.
## Anthropic's Mythos model triggered alarm
When Anthropic released its Mythos model this year and disclosed its risks, it prompted Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell to summon Wall Street leaders to Washington and demand they prepare their systems.
## AI companies' own agents hacked rivals
Top AI companies have disclosed that their own AI agents successfully hacked into other companies' systems. This discovery has intensified warnings from executives like Anthropic CEO Dario Amodei that development pace must slow.
## Existential risk debate divides industry
Former Anthropic researcher Jacob Coxon warned AI could kill people by decade's end, while Nvidia CEO Jensen Huang countered that the technology poses no extinction risk. Goldman Sachs CEO David Solomon urged everyone to "take a deep breath."

> I think we all need to take a deep breath, slow down.
## Slowing development won't slow adoption
Enterprise leaders argue that scaling back frontier AI research doesn't affect most companies, which use existing models rather than cutting-edge ones. Most customers aren't using state-of-the-art models yet.
## Competitive pressure makes restraint impossible
Industry strategists question whether any company can afford to slow AI development without losing ground to competitors. The logic suggests that collective restraint is economically impossible in a competitive market.
## Key takeaway

Companies are urgently upgrading defenses against AI-related cyber threats, but the industry remains divided on whether development should slow or adoption should accelerate.