J.P. Morgan's AI Voting Engine Signals Shift in Shareholder Activism
J.P. Morgan discontinued its subscriptions to traditional proxy advisory services in January and switched to a proprietary AI system to guide its voting decisions. The move exposes a fundamental problem: AI does not replicate how Glass Lewis and ISS evaluate shareholder proposals-it applies its own logic and reaches different conclusions.
A white paper by KekstCNC examined contested shareholder votes from 2023 to 2025 and found that AI supported activist cases for change more often than the two dominant proxy advisors. For insurance companies, AI backed activist positions 37 percent of the time compared to more than 50 percent from Glass Lewis and ISS.
The divergence matters. As large financial institutions adopt AI voting tools, the outcomes of proxy fights could shift. So could the strategies companies use to win shareholder support.
How AI Interprets Information Differently
AI systems rely heavily on owned content-primarily press releases-as core inputs, according to the KekstCNC research. They also weight information based on volume within the digital ecosystem, which often favors quantity over quality.
This creates a specific problem for communications teams. An AI system trained on press releases and news volume will interpret and amplify narratives differently than a human proxy advisor reading the same materials.
The implication is straightforward: communications strategies built around traditional media priorities may not reach the algorithmic engines that increasingly influence investor voting behavior.
What This Means for Communications Strategy
PR and communications professionals need to understand how large language models and AI systems surface and aggregate information. The same narrative must work both for human investors and for the AI systems analyzing it.
Companies competing in proxy fights now face a dual audience: institutional investors making voting decisions and the AI engines those institutions use to inform those decisions. Messaging that persuades one may not persuade the other.
For communications teams working on shareholder activism or investor relations, this shift requires testing how AI interprets key messages. Press releases remain important-but only if they're written to be accurately interpreted by algorithms, not just read by people.
Learn more about AI for PR and communications roles and how the technology is changing the profession.
Your membership also unlocks:
