BankOnIT Insights

CEO Update - Q4 2025

Written by BankOnIT Insights Department | Oct 8, 2025 3:59:44 PM

The AI company Anthropic, best known for its Claude chatbot, did something unexpected. A group of authors had sued the company in 2024, accusing it of using their books to train its AI models without permission. Anthropic recently settled the lawsuit for a landmark $1.5 billion. It agreed to destroy the data used to train its models and set the value of a single infringed book at $3,000. No other AI company being sued by authors has offered anything close.1

Why did Anthropic settle when it appeared they had a good chance at winning their case brought by the authors? It was a strategic move against competitors. At approximately $3,000 per claim, it sets a precedent for similar cases that other AI companies are defending. Other AI companies have far more than $1.5 billion in settlement exposure if the $3,000 per claim is used in the other cases.

How Does This Impact My Bank? As you make decisions about whether and to what extent you want to use automation and AI in your bank part of the conversation will be about which AI firms and tools to utilize. Before you make those decisions, however, you must first conduct an assessment of where your bank is now, where you want it to be in five years, and how best to get there strategically.


Your selection of AI vendors is critical as the next steps require investment of time and resources to integrate the new tools into your existing systems. Selecting the wrong AI vendor or tools could have a significant impact on your institution.

Spending on AI is at epic levels. The artificial-intelligence boom has ushered in one of the costliest building sprees in world history. Over the past three years, leading tech firms have committed more toward AI than it cost to build the interstate highway system, when adjusted for inflation.2

Most AI firms are not currently profitable. Even the largest and most profitable tech companies are not generating a profit from their AI offerings. Expenses are part of the challenge.

Recruiting talent. AI firms compete to recruit the most talented staff and are offering multi-million-dollar pay packages to attract top talent. As an example, a 24-year-old AI researcher received a $250 million pay package over four years from one AI firm.

Data Centers. The costs to build, lease, or buy data centers are enormous. Once built, they then require storage and compute resources, running the latest and most expensive chips, to effectively compete.

Utility Costs. Electricity and connectivity costs for operating AI data centers are significant, with many firms needing to invest in their own power generation or provide utilities with long-term commitments to reduce financial risk to the utility company.

Are we in an AI Bubble? Big losses, low revenue (some AI firms are offering free services), combined with low prices to users, are meant to drive market share.

The Potential Outcome of an AI Bubble. Elimination of free services, significant price increases, and some AI firms may not survive.

Suggestions to Consider. The AI decision should begin with a risk assessment. The assessment should evaluate which AI systems are most compatible with the applications you currently use. The assessment should also consider the risks to your institution if an AI provider is no longer in business, and potential costs if a provider drastically increases their fees.

Involve your Board of Directors. The decision to use AI, how to utilize it, and who to collaborate with is a strategic decision that is best left to the board to make.

Don’t Forget the Regulators. After making a decision, document your reasoning along with a risk assessment of the items considered and include this documentation in board/committee minutes.

Opinion: AI stocks are in a bubble. Why are so many investors refusing to believe it? - AI companies are raising prices and alienating users, and the stocks are overvalued. A well-known Silicon Valley mantra is “Get big or get lost.” The idea is that many markets are essentially winner-take-all — a handful of companies will dominate the market because of daunting economies of scale, allowing the survivors to reap huge profits without fear of competitors entering the market. This sometimes works — think Amazon AMZN — but it often leads to enormous losses as companies prioritize growth over profits. We all remember how the
dot-com bubble ended.3

AI Drives Rise in CEO Impersonator Scams. Cyber crooks are using deepfake voice and videos of top executives to bilk companies out of millions of dollars; it’s ‘No longer a futuristic concept.’4

The Insightful Leader Live: How to Thrive in the Age of AI. In this webinar, Kellogg’s Hatim Rahman, an assistant professor of management and organizations discusses AI in the workforce. “We can choose to use AI to replace as many workers as possible—or we can instead choose to use AI to bolster talent and identify it in underrecognized places.” Rahman further encourages leaders to involve the entire workforce to shape the use of AI within their company.5

IBM Cost of a Data Breach Report 2025. The IBM Cost of a Data Breach Report 2025 was recently released. The global average cost of a data breach for the last year was $4.44 million. The report notes that 20% of the incidents involved shadow AI, the use of AI without employer approval or oversight, and that 16% of breaches involved attackers using AI, primarily in phishing and deepfake attacks. These types of incidents are expected to increase in future years.6

How to Spot North Korean Scammers in the American Workforce. Look for Minions. American tech companies have a serious problem with a despicable twist. The FBI believes thousands of North Koreans have infiltrated the U.S. workforce by assuming the identities of Americans to secure remote jobs. Many of them, investigators have found, are bound by a few defining characteristics: total devotion to Dear Leader Kim Jong Un, a penchant for stealing cryptocurrency, and an obsession with Minions, the cuddly agents of evil from “Despicable Me.”7

October is Cybersecurity Awareness Month. Look for our October Information Security Brief which discusses the history of this declaration and notes four actions everyone can take to strengthen cybersecurity.

1 Wall Street Journal September 8, 2025
2 Wall Street Journal September 25, 2025
3 Marketwatch August 25, 2025
4 Wall Street Journal August 18, 2025
5 The Insightful Leader Live December 7, 2023
6 IBM.com September 24, 2025
7 Wall Street Journal July 22, 2025