Andrew Gordon

1 appearances · 1 as guest · 4 topics

Influence Direction (across 1 analyzed appearances)

Avg Intensity

Moderate 40%

Avg Transparency

Transparent 80%

Top Technique

Direct appeal

Anchoring 1x Confirmation Bias 1x Direct Appeal 1x False Urgency 1x Manufactured Authenticity 1x

Persuasion Dimensions

Implicit Claims

50%

Story Shaping

40%

Emotional Appeal

30%

Call to Action

30%

Engagement Mechanics

20%

ai ethics ai safety llm evaluation machine learning benchmarks

The content aims to advocate for human-centric AI evaluation services, specifically those provided by the guests' company, Prolific.

Viewer Guidance (3 tips)

▶

Question unstated assumptions

Arguments rely on assumptions treated as obvious. Ask what you'd need to already believe for the claims to land.

▶

Consider alternative frames

Information is consistently shaped from one angle. Seek out how other sources present the same facts.

▶

Watch for emotional framing

This content frequently uses emotional appeal. Notice when feelings are being prioritized over evidence.

Questions to Ask Yourself (1)

What would I have to already believe for this argument to make sense?

Implicit Claims — 50%

Machine Learning Street Talk · Discusses the lack of safety metrics in current AI leaderboards and the risks of sensitive user interactions.