Risky Bulletin Episode Summary: "Between Two Nerds: Why Modern Influence Operations Suck"

Release Date: June 16, 2025

Hosted by Tom Uren and Grok, the "Between Two Nerds" episode delves into the efficacy and intricacies of modern influence operations, particularly focusing on the role of artificial intelligence (AI) in shaping these campaigns. Drawing parallels between historical propaganda techniques and contemporary digital strategies, the hosts critically analyze a recent report from Anthropic and explore why current influence operations may fall short of their intended impact.

1. Introduction to AI-Driven Influence Operations

Tom Uren kicks off the discussion by referencing a report from Anthropic released in April, which examines an "influence as a service" platform utilizing Anthropic's language model, Claude, to manage multiple influence campaigns simultaneously.

Tom Uren [00:11]: "I came across this report from Anthropic and came out in April... it talks about an influence as a service platform where they used Anthropic's Claude... to run multiple campaigns at the same time."

Grok expresses skepticism about the sophistication of using AI for influence operations, labeling it as "pedestrian and rudimentary."

Grok [01:08]: "I am rolling my eyes so hard right now."

2. Analysis of Anthropic's Report

The hosts dissect the functionalities of the reported platform:

• Claude's Role: Managing Personas with detailed political alignment guidelines.
• Campaign Activities: Drafting content, evaluating alignment with Personas' viewpoints, generating responses in native languages, and incorporating image generation tools.

Tom Uren posits that the platform appears more as a labor-saving tool or a marketing gimmick rather than a genuinely effective influence mechanism.

Tom Uren [03:56]: "...they've gone all in on getting AI to work. For them, it seems like this is maybe it's a labor saving tool..."

Grok concurs, suggesting that the real value lies in selling the AI-powered influence network rather than its actual utility in influencing opinions.

Grok [04:20]: "It's more of a labor saving device in that some sleazy salesman can scam people into buying their influence agentic thing..."

3. Historical Perspective: WWII Black Propaganda

To highlight the shortcomings of modern AI-driven operations, Grok introduces a comparison with WWII-era British black propaganda efforts aimed at undermining German influence among Bulgarians.

Grok [05:30]: "Yeah. So this, at the time, less sophisticated time, this was called black propaganda..."

Tom Uren and Grok discuss the strategic objectives of the British propaganda, which were clear and purposeful—aimed at demoralizing enemy support and undermining the will to fight.

Grok [07:54]: "They were trying to get the Bulgarians to not support German war aims in general, just to turn against the Germans..."

4. Contrasting Strategic Intent and Execution

The discussion contrasts the intentionality and depth of WWII propaganda with the superficial nature of modern AI-driven campaigns.

Grok emphasizes that historical operations had deeper purposes and strategic objectives, whereas contemporary efforts lack this intentionality.

Grok [22:14]: "The thing that I think is missing is that this political ideology is just the flavor of how you communicate. Whereas the World War II examples, there was these overall objectives... there was more intentionality..."

Tom Uren highlights the absence of a sophisticated decision-making process in the AI-driven platforms, suggesting that AI cannot replace human creativity and strategic thinking essential for effective influence operations.

Tom Uren [21:11]: "...can you replace the human creativity with AI? And I'm like, my guess is no."

5. The Ineffectiveness of Modern Influence Campaigns

The hosts argue that contemporary influence operations, especially those reliant on AI like Claude, are largely ineffective for several reasons:

• Lack of Community Integration: Unlike historical propaganda, modern campaigns often fail to integrate meaningfully within target communities.

• Superficial Engagements: AI-generated content tends to be generic and lacks the nuanced understanding required to genuinely influence opinions.

• Incentive Structures: Both AI providers and influence campaign operators prioritize selling the capability over ensuring its effectiveness.

Grok [18:35]: "It's entirely built around selling this product. Whether you're anthropic selling... or the influence people selling it..."

Tom Uren draws parallels to historical inefficacies, mentioning how certain modern efforts resemble the rudimentary tactics of the Russian IRA in the 2000s, which lacked sophistication and genuine impact.

Tom Uren [25:53]: "Influence campaigns that work are the ones that get influential people and have them promote things like influencer marketing."

6. The Oversaturation and Distrust of Disinformation

Grok discusses the current landscape where disinformation and influence campaigns are so pervasive that they have lost credibility and effectiveness. The overuse of such tactics leads to public skepticism, further diminishing their impact.

Grok [26:11]: "Influencer marketing... doesn't work so well these days because it's being exploited to death and so people don't trust it as much."

Tom Uren echoes this sentiment, noting that without genuine engagement and trust within communities, influence campaigns fail to resonate.

Tom Uren [26:26]: "If it's just someone random online who could be a bot, then it's I guess not going to be that engaging or effective."

7. Historical Lessons and Future Outlook

The episode concludes with reflections on how historical propaganda was once seen as a groundbreaking form of warfare, only to be later understood with greater nuance. Similarly, modern digital propaganda is deemed overhyped and not as impactful as initially feared.

Grok [27:18]: "It's been rediscovered for the Internet, like, oh, my God, propaganda... And then it turns out that the next war isn't fought entirely with propaganda over the Internet."

Tom Uren humorously contemplates the longevity of their discussions on warfare and propaganda.

Tom Uren [30:03]: "Yeah, I doubt we'll be podcasting then."

Concluding Insights

The hosts assert that modern influence operations, particularly those leveraging AI without strategic depth and community integration, are largely ineffective. Drawing from historical examples, they emphasize the importance of intentionality, strategic objectives, and genuine engagement in successful propaganda and influence endeavors. The episode serves as a critique of superficial digital influence tactics, advocating for more thoughtful and integrated approaches to information warfare.

Notable Quotes:

• Grok [03:09]: "It's like somewhat like it's, you know, used a fork. Use the fork to eat a steak. Use the fork to eat broccoli..."

• Tom Uren [10:00]: "It's just, well, so anthropic rights, you know, four distinct campaigns. The one I like most is, number two, promoting the United Arab Emirates as the superior business environment while criticizing EU regulatory frameworks."

• Grok [14:42]: "And far away, your sons are going to die in France for Hitler... it's so bad."

• Grok [26:11]: "Influencer marketing... doesn't work so well these days because it's being exploited to death and so people don't trust it as much."

This episode of "Risky Bulletin" offers a critical examination of the current state of influence operations, juxtaposing historical strategies with modern AI-driven methods to underscore the complexities and challenges inherent in shaping public opinion through digital means.