AI News of the Month | July 2025

While I was trying to enjoy some rest during my Lisbon "workation," the AI industry certainly didn't take a break.

July brought significant AI tool launches, critical regulatory developments, and notable insights from executives at leading AI companies.

Here's what's most relevant for the finance teams

What’s New in AI

1. Anthropic Launches Claude for Financial Services

Anthropic made a major move this month, launching Claude for Financial Services—a specialized version of Claude specifically tailored for finance, insurance, and banking industries. This new platform integrates seamlessly with key tools such as Databricks, Snowflake, Morningstar, and S&P Global, providing robust support for market research, investment modeling, automated compliance, and enhanced risk management. Claude Opus 4 notably achieved an 83% accuracy rate on complex Excel modeling benchmarks, highlighting its strength as a powerful analyst tool.

So What:
I've long emphasized that general finance is underserved by AI—but financial services (banking, insurance, fintech) are an exception, being particularly ripe for AI-driven innovation due to the sheer volume of data involved. Among current large language models, Claude is already recognized as one of the safest, especially in terms of data privacy. This makes Claude for Financial Services an excellent fit. Even if you're not directly in fintech, insurance, or banking, you’ll likely feel the positive effects—potentially seeing improved customer experiences and reduced service costs. Regardless, it's a great example of industry-specific AI done right.

2. OpenAI Introduces ChatGPT Agent—With a Warning

OpenAI unveiled its much-anticipated ChatGPT Agent, designed to autonomously carry out multi-step tasks such as populating spreadsheets, conducting detailed web research, and generating presentations. However, CEO Sam Altman clearly cautioned against using this agent for high-stakes financial tasks without strict oversight, warning that it remains experimental and prone to errors or misuse.

So What:
AI agents are advancing quickly, but their reliability for critical finance tasks is still limited.

What to Do:
Experiment cautiously. Begin by deploying ChatGPT Agent in clearly defined, low-risk workflows. Maintain strong human oversight, especially in these initial stages, to ensure accuracy and avoid potential problems.

From my own brief experiments, I haven't been particularly impressed—so far, the agent struggles even with basic tasks. But remember, ChatGPT itself started with many errors and limited capabilities. Today, it's incredibly powerful. Agent-based AI is undoubtedly part of the future, so it's wise to start exploring now, even if cautiously. Just ensure you have strong safety measures and robust controls in place.

3. ChatGPT Conversations Lack Legal Confidentiality

In an important clarification, OpenAI CEO Sam Altman confirmed that ChatGPT conversations carry no legal privilege, such as attorney-client or doctor-patient confidentiality. This means your ChatGPT logs could potentially be subpoenaed and used as evidence in legal proceedings—even data you've deleted within the 30-day retention window.

So What:
This development has significant implications for data privacy, compliance, and litigation. Honestly, this is also scary!

What to Do:
Immediately update your internal AI usage policies. Avoid using free versions of ChatGPT for tasks involving sensitive information, and clearly communicate this confidentiality limitation to all employees.

As someone already cautious in my AI practices, this news makes me even more conservative about how I interact with public AI tools, and I recommend you adopt a similar approach.

4. Regulatory & Governance Updates Roundup

July brought several key regulatory developments that impact financial services:

• U.S. Senate Preserves State-Level AI Regulation:
  The Senate rejected a bill that aimed to stop state-level AI regulation for ten years, meaning businesses face continued complexity from state-specific compliance rules.
  

• America’s AI Action Plan:
  The federal government revealed a strategy prioritizing deregulation, state-level incentives, and enhanced AI research infrastructure. Financial institutions may see deregulation in some areas but increased oversight in others, particularly around exports and high-risk applications.
  

• Unleashing AI Innovation in Financial Services Act:
  Newly introduced legislation proposes regulatory sandboxes, allowing financial institutions to safely pilot innovative AI solutions under regulatory supervision.
  

• EU's GPAI Code of Practice:
  The EU released guidelines on transparency, data sourcing, risk assessment, and security for general-purpose AI tools such as ChatGPT and Claude. Compliance is mandatory, with potential fines reaching up to 7% of global turnover.
  

• Council of Europe AI Framework Convention:
  Over 50 countries signed this treaty, committing to transparency, non-discrimination, and robust AI governance aligned with human rights principles.
  

So What:
The regulatory landscape is rapidly evolving, and finance executives and compliance teams need to stay proactive and agile.

After a month of intensive focus on AI regulations, I can confidently say this area remains incredibly dynamic, with a challenging balance between innovation and oversight yet to be established. I'll continue closely monitoring this evolving landscape.

What to Do:

• Assess your organization's exposure to state, federal, and international AI regulations.

• Proactively engage with regulatory sandboxes and similar pilot programs.

• Align your organization's AI governance with EU and international standards to ensure compliance.

WRAP-UP

July has reinforced the accelerating pace of AI adoption alongside increased complexity in regulation and confidentiality—particularly within financial services. Navigating this landscape successfully will depend heavily on cautious but proactive governance. If your organization is grappling with these opportunities and challenges, stay tuned—I'll continue to highlight what's most critical.

Closing Thoughts

July was certainly not a vacation month for AI and finance—and it doesn’t look like August will be any quieter. Some of this month’s developments feel like they will shape the future of AI, and in August, we’re expecting the release of ChatGPT 5.0 along with other significant innovations.

The regulatory landscape is becoming increasingly complex, and it’s fascinating to watch how different countries and regions are shaping their attitudes and approaches to AI governance.

In any case—stay connected, I’ll keep sharing the updates that matter most for CFOs and finance leaders navigating this fast‑changing space.

We Want Your Feedback!

This newsletter is for you, and we want to make it as valuable as possible. Please reply to this email with your questions, comments, or topics you'd like to see covered in future issues. Your input shapes our content!

Want to dive deeper into balanced AI adoption for your finance team? Or do you want to hire an AI-powered CFO? Book a consultation! 

Did you find this newsletter helpful? Forward it to a colleague who might benefit!

Until next Tuesday, keep balancing!

Anna Tiomina 
AI-Powered CFO