Skip to main content

​
Foundational risks of AI in society

Time: 9:30 AM to 10:10 AM
This week you built a robot that sees, hears, speaks, and acts on LLM instructions. Today starts with a direct question: what could go wrong?

​
Bias and hallucination

LLMs confidently generate incorrect information. The model does not know it is wrong. When these systems make decisions in hiring, healthcare, or law enforcement, the consequences are real.

​
Surveillance and misuse

The same camera and VLM stack you built this week could be used to identify faces, track people, and classify behavior at scale. The question is not whether the technology can do this. It already can. The question is: who owns that capability, and who is accountable?

​
Concentration of power

The most capable models are owned by a handful of companies. What happens when critical infrastructure depends on a private API that can be shut off, rate-limited, or priced out of reach?

​
Autonomous weapons

The architecture you built this week, perceive, reason, act, is the same foundation used in lethal autonomous weapon systems (LAWS). This is not theoretical. The international community is actively debating whether machines should make targeting decisions without human approval.

​
The dual-use problem

The same research that enables a helpful robot enables a harmful one. There is no technical separation between the two. The difference is intent and governance.
Ground this in what you built: “We just built a robot that can identify objects and move toward them. Now imagine that at scale, with a weapon attached.” These are real stakes, not abstract ones.