Jakarta - Concerns about the latest generation of artificial intelligence are beginning to be felt in the global financial sector. Goldman Sachs CEO David Solomon admitted to being "hyper-aware" of the risks posed by the latest AI model called Mythos, developed by Anthropic.
This statement comes amid growing concerns that AI is now not only a tool - but also a potential serious threat to cybersecurity.
AI that Can "Break" the System
The Mythos model, part of Anthropic's Claude family, is said to have an extraordinary ability to find and exploit software security gaps.
In his statement, Anthropic even admitted that this model is able to surpass most humans in detecting system vulnerabilities.
"AI models have reached a level of coding ability where they can surpass almost all humans in finding and exploiting software vulnerabilities," Anthropic wrote in its official blog.
What does it mean? The world is facing a new possibility: AI that can think like a hacker, but with the speed of a machine.
Goldman Sachs Not Staying Silent
Facing this potential threat, Goldman Sachs chose a proactive approach. Solomon emphasized that his party already had access to the model and worked directly with Anthropic and security vendors to understand and utilize it.
"We recognize the capabilities of Mythos... and we are working closely with Anthropic and our security partners to leverage this frontier technology," Solomon said.
He added that investment in the cybersecurity sector is now being accelerated, as part of a strategy to strengthen the resilience of the bank's digital infrastructure.
Alarm from Global Regulators
This concern does not only come from the private sector. The AI Security Institute (AISI) calls Mythos a "major leap" in the potential of cyber threats.
In their tests, the model was able to complete a complex cyberattack simulation with 32 steps - something that would normally take human professionals days - and was successful in several trials.
More worrying, this AI is considered capable of carrying out attacks autonomously against systems with weak defenses.
However, AISI also noted that the ability to deal with systems with strong defenses is still uncertain.
This situation triggered a quick response from regulators. US Treasury Secretary Scott Bessent has even gathered the bosses of major banks to discuss the risks posed by this AI.
In the UK, the Cross Market Operational Resilience Group is scheduled to hold a similar meeting in the near future, involving authorities such as the Bank of England and the Financial Conduct Authority.
This development confirms one reality: AI is now not only a productivity tool, but also a potential cyber weapon.
For financial institutions, the stakes are very high. Even a small disruption can have a systemic impact on economic stability.
The English, Chinese, Japanese, Arabic, and French versions are automatically generated by the AI. So there may still be inaccuracies in translating, please always see Indonesian as our main language. (system supported by DigitalSiber.id)
