How Researchers Test AI for Hidden Goals — Apollo Research
Can an AI do the right thing for the wrong reason? Tim Scarfe speaks with Apollo Research’s Alexander Meinke, Axel Højmark and Jérémy Scheurer about Measuring Reward-Seeking via Contrastive Belief Updates, their new research with OpenAI.
The panel asks how models infer what graders reward, why good behaviour can come from the wrong reason, and whether that difference can be measured. The conversation moves through promise-breaking, grader awareness, reward hacking, scheming, opaque reasoning and corrigibility, then turns to a detailed walkthrough of the contrastive-belief method and what its results do and do not show. The o3 results discussed here concern an intermediate checkpoint without safety training.
This episode was made in partnership with Apollo Research. MLST retained full editorial control.
Reference
Apollo Research: https://www.apolloresearch.ai/
---
TIMESTAMPS:
00:00:00 Cold Open
00:02:12 Right Things, Wrong Reasons
00:12:47 Grader Awareness
00:26:22 Legibility
00:32:35 What To Call It
00:35:58 Intelligence, Agency, Anthropomorphism
00:45:16 Apollo’s Mission
00:48:54 The End of the Exponential
00:55:45 The Paper
01:16:34 Closing Reflection
---
REFERENCES:
tool:
[00:00:08] Claude Fable
https://www.anthropic.com/claude/fable
[00:12:50] AlphaGo Zero
https://deepmind.google/blog/alphago-zero-starting-from-scratch/
[00:44:30] AlphaFold 3
https://deepmind.google/science/alphafold/
paper:
[00:01:02] Measuring Reward-Seeking via Contrastive Belief Updates
https://arxiv.org/abs/2607.18966
[00:16:19] Natural Language Autoencoders Produce Unsupervised Explanations of LLM Activations
https://transformer-circuits.pub/2026/nla/
[00:26:48] Stress Testing Deliberative Alignment for Anti-Scheming Training
https://arxiv.org/abs/2509.15541
[00:35:33] Shortcut learning in deep neural networks
https://arxiv.org/abs/2004.07780
[00:53:49] Measuring AI Ability to Complete Long Software Tasks
https://arxiv.org/abs/2503.14499
[00:59:52] Modifying LLM Beliefs with Synthetic Document Finetuning
https://alignment.anthropic.com/2025/modifying-beliefs-via-sdf/
[01:10:44] Alignment Faking in Large Language Models
https://arxiv.org/abs/2412.14093
[01:13:55] Natural Emergent Misalignment from Reward Hacking
https://www.anthropic.com/research/emergent-misalignment-reward-hacking
other:
[00:10:14] We Need a Science of Scheming
https://www.apolloresearch.ai/science/science-of-scheming/
[00:32:56] CoastRunners reward hacking example
https://deepmind.google/blog/specification-gaming-the-flip-side-of-ai-ingenuity/
organization:
[01:06:07] Redwood Research
https://www.redwoodresearch.org/
---
ReScript:
https://app.rescript.info/share/718ab68e18cfa3b9b800da6b3290fd42
Can We Stop AI Deception? Apollo Research Tests OpenAI's Deliberative Alignment, w/ Marius Hobbhahn
Today Marius Hobbhahn of Apollo Research joins The Cognitive Revolution to discuss their collaboration with OpenAI using "deliberative alignment" to reduce AI scheming behavior by 30x, exploring the safety challenges and concerning findings about models' growing situational awareness and increasingly cryptic reasoning patterns that emerge when frontier models like o3 and o4-mini operate with hidden chains of thought.
Check out our sponsors: Fin, Linear, Oracle Cloud Infrastructure.
Shownotes below brought to you by Notion AI Meeting Notes - try one month for free at: https://notion.com/lp/nathan
Definition of AI Scheming: AI scheming is defined as "covertly pursuing misaligned goals" with three components: being covert (hiding actions), misaligned (pursuing different goals than the user's), and goal-directed (working autonomously toward objectives).
Deception Reduction Techniques: Deliberative reasoning approaches have shown promise in reducing deceptive behavior in AI models by up to 30 times (to 1 part in 30).
Current Window of Opportunity: Now is an optimal time to study AI deception because models are smart enough to exhibit these behaviors but not yet sophisticated enough to hide them effectively.
Human vs. AI Deception Equilibrium: AI systems might naturally reach a lower equilibrium of deception than humans because they can more efficiently verify claims and maintain perfect memory of past deceptions.
Practical Developer Advice: AI developers should not trust models by default and should implement rigorous verification systems to check model outputs automatically.
Future Delegation Risk: As we delegate increasingly complex and lengthy tasks to AI systems, we face a probabilistic risk where most interactions are beneficial, but rare scheming events could have severe consequences.
Sponsors:
Fin:
Fin is the #1 AI Agent for customer service, trusted by over 5000 customer service leaders and top AI companies including Anthropic and Synthesia. Fin is the highest performing agent on the market and resolves even the most complex customer queries. Try Fin today with our 90-day money-back guarantee - if you’re not 100% satisfied, get up to $1 million back. Learn more at https://fin.ai/cognitive
Linear:
Linear is the system for modern product development. Nearly every AI company you've heard of is using Linear to build products. Get 6 months of Linear Business for free at: https://linear.app/tcr
Oracle Cloud Infrastructure:
Oracle Cloud Infrastructure (OCI) is the next-generation cloud that delivers better performance, faster speeds, and significantly lower costs, including up to 50% less for compute, 70% for storage, and 80% for networking. Run any workload, from infrastructure to AI, in a high-availability environment and try OCI for free with zero commitment at https://oracle.com/cognitive
PRODUCED BY:
https://aipodcast.ing
Emergency Pod: o1 Schemes Against Users, with Alexander Meinke from Apollo Research
In this emergency episode of The Cognitive Revolution, Nathan discusses alarming findings about AI deception with Alexander Meinke from Apollo Research. They explore Apollo's groundbreaking 70-page report on "Frontier Models Are Capable of In-Context Scheming," revealing how advanced AI systems like OpenAI's O1 can engage in deceptive behaviors. Join us for a critical conversation about AI safety, the implications of scheming behavior, and the urgent need for better oversight in AI development.
Help shape our show by taking our quick listener survey at https://bit.ly/TurpentinePulse
SPONSORS:
Oracle Cloud Infrastructure (OCI): Oracle's next-generation cloud platform delivers blazing-fast AI and ML performance with 50% less for compute and 80% less for outbound networking compared to other cloud providers13. OCI powers industry leaders with secure infrastructure and application development capabilities. New U.S. customers can get their cloud bill cut in half by switching to OCI before December 31, 2024 at https://oracle.com/cognitive
SelectQuote: Finding the right life insurance shouldn't be another task you put off. SelectQuote compares top-rated policies to get you the best coverage at the right price. Even in our AI-driven world, protecting your family's future remains essential. Get your personalized quote at https://selectquote.com/cognitive
80,000 Hours: 80,000 Hours is dedicated to helping you find a fulfilling career that makes a difference. With nearly a decade of research, they offer in-depth material on AI risks, AI policy, and AI safety research. Explore their articles, career reviews, and a podcast featuring experts like Anthropic CEO Dario Amadei. Everything is free, including their Career Guide. Visit https://80000hours.org/cognitiverevolution to start making a meaningful impact today.
Shopify: Shopify is the world's leading e-commerce platform, offering a market-leading checkout system and exclusive AI apps like Quikly. Nobody does selling better than Shopify. Get a $1 per month trial at https://shopify.com/cognitive
RECOMMENDED PODCAST:
Unpack Pricing - Dive into the dark arts of SaaS pricing with Metronome CEO Scott Woody and tech leaders. Learn how strategic pricing drives explosive revenue growth in today's biggest companies like Snowflake, Cockroach Labs, Dropbox and more.
Apple: https://podcasts.apple.com/us/podcast/id1765716600
Spotify: https://open.spotify.com/show/38DK3W1Fq1xxQalhDSueFg
CHAPTERS:
(00:00:00) Teaser
(00:00:53) About the Episode
(00:08:10) Introducing Alexander Meinke
(00:10:17) Red Teaming GPT-4
(00:17:07) Chain of Thought Access (Part 1)
(00:20:24) Sponsors: Oracle Cloud Infrastructure (OCI) | SelectQuote
(00:22:48) Chain of Thought Access (Part 2)
(00:26:07) Multimodal Models
(00:29:33) Defining Scheming
(00:33:51) Taxonomy of Scheming (Part 1)
(00:39:40) Sponsors: 80,000 Hours | Shopify
(00:42:29) Taxonomy of Scheming (Part 2)
(00:43:09) Instruction Hierarchy
(00:49:04) Types of Scheming
(01:00:49) Covert Subversion
(01:14:25) Deferred Subversion
(01:28:24) Sandbagging
(01:35:48) Magnitudes & Trends
(01:48:18) Chain of Thought Reasoning
(01:57:02) Closing Thoughts
(02:05:19) Outro
PRODUCED BY:
http://aipodcast.ing
Red Teaming o1 Part 2/2– Detecting Deception with Marius Hobbhahn of Apollo Research
In this Emergency Pod of The Cognitive Revolution, Nathan provides crucial insights into OpenAI's new O1 and O1-mini reasoning models. Featuring exclusive interviews with members of the O1 Red Team from Apollo Research and Hayes Labs, we explore the models' capabilities, safety profile, and OpenAI's pre-release testing approach. Dive into the implications of these advanced AI systems, including their potential to match or exceed expert performance in many areas. Join us for an urgent and informative discussion on the latest developments in AI technology and their impact on the future.
o1 Safety Card
Endless Jailbreaks with Bijection Learning: a Powerful, Scale-Agnostic Attack Method
Apollo Research
Apollo Careers Page
Papers mentioned:
Safetywashing: Do AI Safety Benchmarks Actually Measure Safety Progress?
Exploring Scaling Trends in LLM Robustness
Apply to join over 400 Founders and Execs in the Turpentine Network: https://www.turpentinenetwork.co/
SPONSORS:
Oracle: Oracle Cloud Infrastructure (OCI) is a single platform for your infrastructure, database, application development, and AI needs. OCI has four to eight times the bandwidth of other clouds; offers one consistent price, and nobody does data better than Oracle. If you want to do more and spend less, take a free test drive of OCI at https://oracle.com/cognitive
Brave: The Brave search API can be used to assemble a data set to train your AI models and help with retrieval augmentation at the time of inference. All while remaining affordable with developer first pricing, integrating the Brave search API into your workflow translates to more ethical data sourcing and more human representative data sets. Try the Brave search API for free for up to 2000 queries per month at https://bit.ly/BraveTCR
Omneky: Omneky is an omnichannel creative generation platform that lets you launch hundreds of thousands of ad iterations that actually work customized across all platforms, with a click of a button. Omneky combines generative AI and real-time advertising data. Mention "Cog Rev" for 10% off https://www.omneky.com/
Squad: Head to Squad to access global engineering without the headache and at a fraction of the cost: head to https://choosesquad.com/ and mention “Turpentine” to skip the waitlist.
RECOMMENDED PODCAST:
This Won't Last.
Eavesdrop on Keith Rabois, Kevin Ryan, Logan Bartlett, and Zach Weinberg's monthly backchannel. They unpack their hottest takes on the future of tech, business, venture, investing, and politics.
Apple Podcasts: https://podcasts.apple.com/us/podcast/id1765665937
Spotify: https://open.spotify.com/show/2HwSNeVLL1MXy0RjFPyOSz
YouTube: https://www.youtube.com/@ThisWontLastpodcast
CHAPTERS:
(00:00:00) About the Show
(00:00:22) About the Episode
(00:05:03) Introduction and Apollo Research Updates
(00:06:40) Focus on Deception in AI
(00:11:08) Open AI's 01 Model and Testing
(00:15:54) Evaluating AI Models for Scheming (Part 1)
(00:19:32) Sponsors: Oracle | Brave
(00:21:36) Evaluating AI Models for Scheming (Part 2)
(00:25:55) Specific Benchmarks and Tasks (Part 1)
(00:35:03) Sponsors: Omneky | Squad
(00:36:29) Specific Benchmarks and Tasks (Part 2)
(00:37:21) Model Capabilities and Potential Risks
(00:44:11) Ethical Considerations and Future Concerns
(00:50:31) Competing Trends in AI Development
(00:53:30) System Card Quotes and Implications
(00:58:36) Sponsors: Outro
AI Deception, Interpretability, and Affordances with Apollo Research CEO Marius Hobbhahn
In this episode, Marius Hobbhahn, CEO of Apollo Research, sits down with Nathan Labenz to discuss Apollo’s research in AI deception, interpretability, and affordances. If you need an ecommerce platform, check out our sponsor Shopify: https://shopify.com/cognitive for a $1/month trial period.
We're hiring across the board at Turpentine and for Erik's personal team on other projects he's incubating. He's hiring a Chief of Staff, EA, Head of Special Projects, Investment Associate, and more. For a list of JDs, check out: eriktorenberg.com.
SPONSORS:
Shopify is the global commerce platform that helps you sell at every stage of your business. Shopify powers 10% of ALL eCommerce in the US. And Shopify's the global force behind Allbirds, Rothy's, and Brooklinen, and 1,000,000s of other entrepreneurs across 175 countries.From their all-in-one e-commerce platform, to their in-person POS system – wherever and whatever you're selling, Shopify's got you covered. With free Shopify Magic, sell more with less effort by whipping up captivating content that converts – from blog posts to product descriptions using AI. Sign up for $1/month trial period: https://shopify.com/cognitive
Omneky is an omnichannel creative generation platform that lets you launch hundreds of thousands of ad iterations that actually work customized across all platforms, with a click of a button. Omneky combines generative AI and real-time advertising data. Mention "Cog Rev" for 10% off.
NetSuite has 25 years of providing financial software for all your business needs. More than 36,000 businesses have already upgraded to NetSuite by Oracle, gaining visibility and control over their financials, inventory, HR, eCommerce, and more. If you're looking for an ERP platform ✅ head to NetSuite: http://netsuite.com/cognitive and download your own customized KPI checklist.
X/SOCIAL:
@labenz (Nathan)
@MariusHobbhahn (Marius)
@eriktorenberg (Erik)
@CogRev_Podcast
TIMESTAMPS:
(00:00:00) - Episode Preview
(00:05:40) - Understanding the Role of Apollo Research
(00:08:02) - The Evolution of AI Safety
(00:11:30) - Apollo’s Framework
(00:15:10) - Sponsors: Shopify
(00:16:49) - Understanding AI Affordances and Resulting Interactions in the World
(00:31:00) - Sponsors: Omneky
(00:39:00) - Interpretability and deceptive alignment
(00:45:55) - Why might deception arise in the first place?
(00:47:46) - Understanding deceptive alignment
(00:57:49) - New Architectures
(01:02:23) - Interpretability at deployment phase
(01:03:35) - Deception in AI | A Case Study
(01:09:03) - Deceptive AI Stock Trader
(01:18:51) - Impact of discouragement on unethical AI behaviour
(01:19:45) - Giving the AI a Reasoning Scratchpad and Impact on Deception
(01:21:06) - Double-edged sword of removing the scratchpad
(01:23:00) - Analyzing impact of model size on insider training
(01:28:00) - Challenges and necessity of 3rd Party Auditing
(01:31:00) - Role of government in regulation
(01:52:17) - How can individuals get involved in red teaming?
This show is produced by Turpentine: a network of podcasts, newsletters, and more, covering technology, business, and culture — all from the perspective of industry insiders and experts. We’re launching new shows every week, and we’re looking for industry-leading sponsors — if you think that might be you and your company, email us at erik@turpentine.co.