Harvard Launches AI Safety Graduate Course with OpenAI and Anthropic Speakers

Harvard is offering a graduate AI safety course this fall, featuring researchers from OpenAI and Anthropic and free lecture recordings on YouTube.
Harvard is running a graduate AI safety course this fall, and the list of invited specialists is impressive.
From OpenAI, chief scientist Jakub Pachocki and co-founder John Schulman will speak, and a Q&A session with Sam Altman is also planned.
From Anthropic — Nicholas Carlini and Jack Lindsey.
The program covers:
- RLHF and reinforcement learning for reasoning models;
- reward hacking, when models find ways to fool evaluators;
- monitoring chains of reasoning;
- interpretability, i.e., studying the internal structure of models;
- cyber and biological AI risks;
- AI alignment when AI helps create the next generation of AI;
- model specifications, open weights, regulation and AI's impact on the labor market.
Lectures can be watched for free on YouTube. More details are available in the course program.
Why it matters
AnalysisThe course shows that AI safety is entering mainstream academic curricula and attracting leading researchers from frontier labs. It helps readers understand RLHF, reward hacking, interpretability and governance debates around AI.
Discuss in community
Share your questions and insights with developers
Related in Ecosystem
Practical guides, core terminology, and agent skills
Anthropic goes public with shares trading near $100
Anthropic completed an $81 billion IPO and raised another $26 billion through a private placement to large investors, with demand exceeding supply sevenfold.
Anthropic mathematician claims possible proof of Riemann hypothesis with Claude
Levent Alpoge says collaborative work with Claude may have led to a proof of the Riemann hypothesis.
Claude Projects (Workspaces with Custom Knowledge Base)
A workspace organization tool in Claude by Anthropic. It allows the creation of dedicated workspaces for each task, uploading up to 200,000 tokens of internal documentation, and sharing them with colleagues.
Claude Haiku (Fast Compact Models by Anthropic)
The most compact and fastest model in the Anthropic lineup, designed for lightning-fast responses, mass text classification, data extraction, and agent routing at minimal cost.