Goldman Sachs chief executive David Solomon has said he is “hyper-aware” of the capabilities of Anthropic’s Mythos AI model, as the bank works closely with the company after it issued warnings about the cybersecurity risks the system could pose.
The remarks underline how major financial institutions are balancing the promise of fast-moving artificial intelligence tools with the need to guard against misuse by hackers and other malicious actors.
Goldman Sachs has been monitoring the rapid progress of artificial intelligence, including large language models, as part of broader efforts to strengthen its defences against cyber threats. The bank is one of the US financial firms closely tracking how these systems might affect security, operations and risk management.
According to the source, Goldman Sachs has the Claude model and is collaborating with Anthropic to improve cyber protection. The relationship reflects growing interest across the banking sector in using AI systems while also preparing for the risks they may introduce.
Solomon’s comments suggest that the firm is not only interested in the capabilities of advanced AI, but also alert to the possibility that such tools could be used in ways that increase vulnerability. That concern has become more urgent as models become more powerful and more widely available.
The source does not detail any specific incident involving Mythos AI, but it makes clear that Anthropic has warned about the cybersecurity risk the model may present. Goldman Sachs’s response has been to stay closely engaged with the company while continuing to monitor developments in the wider AI landscape.
For large banks, the challenge is increasingly about keeping pace with technology that can both strengthen and threaten security systems. Goldman Sachs’s approach, as described in the source, reflects a wider push in the industry to understand the risks of artificial intelligence before they become a bigger problem.
