The fiduciary duty of loyalty, as delineated in Section 78(3) of the Restatement (Third) of Trusts, imposes a stringent standard on fiduciaries, including plan sponsors and investment fiduciaries. The language “knew or should have known” underscores the expectation that fiduciaries remain vigilant and informed about their decisions and the tools they employ, such as artificial intelligence (AI).

In the retirement plan context (e.g., ERISA fiduciaries, plan sponsors, investment committees), courts and regulators often look to the Restatement for interpretive guidance on duty of prudence, loyalty, and monitoring obligations.

Under section 79(3), fiduciaries face exposure if they:

1. Knew of a breach (actual knowledge), or
2. Should have known (constructive knowledge based on what a reasonably prudent fiduciary would have known under the circumstances).

The “should have known” standard is objective and evolves with available tools, practices, and information in the fiduciary community.

A prudent fiduciary response requires

• 1. Demonstrated awareness of AI-enabled capabilities,
• 2. Deliberate governance decisions about their use, and
• 3. Robust documentation of the rationale behind those decisio

📌 Key Fiduciary Duties Under Section 78(3)

Section 78(3) articulates that a fiduciary must act with the utmost loyalty, avoiding conflicts of interest and placing the beneficiaries’ interests above all else. The “knew or should have known” standard implies that fiduciaries are expected to be aware of potential risks and to act proactively to mitigate them.

The “knew or should have known” standard is also repeated in ERISA scetion 404(a). Plan sponsors cannot legitimately claim to be surprised by the standard In the retirement plan context (e.g., ERISA fiduciaries, plan sponsors, investment committees), courts and regulators often look to the Restatement for interpretive guidance on duty of prudence, loyalty, and monitoring obligations. This is in accordance with the the Supreme Court’s direction in the landmark Tibble¹ decision – look to the Restatement of Trusts when fiduciary issues ar involved

The “knew or should have known” standard emphasizes the need for fiduciaries to be proactive and informed. While AI offers significant advantages in financial decision-making, fiduciaries must exercise caution, ensuring that their use of AI aligns with their duty of loyalty and care. By implementing due diligence, oversight, and expert consultation, fiduciaries can navigate the complexities of AI integration while upholding their fiduciary responsibilities.

The proliferation of AI-based research, analytics, and monitoring tools (e.g., for investment due diligence, ESG risk screening, fee benchmarking, and market anomaly detection) may raise the baseline of reasonable prudence.

A fiduciary who fails to use—or at least consider—the insights available from these technologies could be deemed to have constructive knowledge of risks or issues that AI tools would have revealed.

AI and the Evolving Standard of Care

The “knew or should have known” standard emphasizes the need for fiduciaries to be proactive and informed. While AI offers significant advantages in financial decision-making, fiduciaries must exercise caution, ensuring that their use of AI aligns with their duty of loyalty and care. By implementing due diligence, oversight, and expert consultation, fiduciaries can navigate the complexities of AI integration while upholding their fiduciary responsibilities.

Key Implication:

As AI becomes a widely available tool, courts and regulators may infer that a prudent fiduciary should have known what a commercially reasonable AI system could have shown.

Under Section 78(3), fiduciaries can be liable not only for what they knew but also for what they should have known—as the standard of what they “should have known” evolves with technology.

The increasing accessibility and sophistication of AI-driven fiduciary tools raise the bar for prudence and oversight. Plan sponsors and fiduciaries who fail to integrate or reasonably evaluate such tools risk constructive knowledge liability and co-fiduciary exposure.

Courts and regulators are applying increasingly technologically informed standards of prudence. As fiduciary tools advance, the threshold for what a fiduciary “should have known” correspondingly increases. AI-enabled analytics are no longer experimental—they are rapidly becoming an industry norm.

Under Section 78(3), the constructive knowledge standard may expand as AI tools make previously opaque information discoverable. Fiduciaries can no longer credibly claim ignorance of risks that automated tools routinely identify in peer institutions.

Regulators (e.g., DOL, SEC) and plaintiffs’ counsel increasingly reference data analytics capabilities as part of fiduciary process evaluation. Fiduciaries who cannot demonstrate awareness of these tools face heightened litigation vulnerability and discovery risk.

The fiduciary duty of prudence is dynamic. As AI and data-driven research redefine what fiduciaries can reasonably know, Section 78(3) increases the risk that inaction or failure to adopt available tools will be seen as a breach of fiduciary duty.

The “should have known” standard is objective and evolves with available tools, practices, and information in the fiduciary community.

Going Forward
The likelihood that the courts may adopt a AI-generated breakeven analysis of annuities going forward is as much a matter of simplicity as a seach for equiatable decisions. The introduction of AI into questions of fiduciary prudence and fiducairy litigation zllows for quick and simple analyses, where changes in input data can be easily made, allowing for easier targeting of causation factors.My own experience in using AI to prepare annuity breakeven analysis have produced excellent results, allowing for properly factoring in both present value and mortality risk in the fiduciary prudence analysis. The ease and cost-efficiency of AI-generated fiduciary prudence analytics makes the inclusion of same in a plan’s process a must. An example of a sample prompt that I have used with success is

Prepare a breakeven analysis, brokern into annual intervals, factoring in both present value and mortality risk, on a $$$$$$ immediatew annuity for a 65 year-old male/female retiring at age X, asuming a normal life expectancy.

ERISA plaintifff attorneys are already using AI to develop cases and to design trial strategies. Plan sponsors can also use AI to create screens and interview strategies for vendos and plan advisers.

Notes
1. Tibble v. Edison International, 135 S. Ct. 1823, 1828 (2015).

© Copyright 2025 InvestSense, LLC. All rights reserved.

This article is for informational purposes only and is neither designed nor intended to provide legal, investment, or other professional advice since such advice always requires consideration of individual circumstances. If legal, investment, or other professional assistance is needed, the services of an attorney or other professional advisor should be sought.