Scores by kind of claim
Each kind of claim is scored on its own, out of 100: the share that held up, with the range it likely falls in. Fewer resolved claims give a wider range.
-
Factual claims
Statements about the present or the past, checked against the record: did they hold up?
72 65 79 -
Predictions
Statements about events outside the person's control, checked once they fell due: did they come true?
66 54 78 -
Promises
Commitments to the person's or organisation's own future action, with something to deliver: was it delivered?
Insufficient record: 1 promise resolved, 5 needed for a score.
All types combined: 71 (likely 64 to 77, high confidence), from 79 claims: 60 factual claims, 19 predictions. It mixes kinds of claim in different proportions for different people, so compare people kind by kind.
Balaji Srinivasan's claims
37 claims
“Killer AI is already here, and it’s called drones.”
Balaji Srinivasan, author of his own blog post 3 August 2025 AI is polytheistic, not monotheistic (balajis.com)
Killer AI already exists in the form of drones, and every country is pursuing it.
Mostly true True in substance, wrong in a detail.
The second judge would not resolve this claim; the verdict shown is the first judge's.
What would count as true or false
The claim has two parts: (1) lethal drones with AI or autonomous targeting existed by August 2025; (2) every country, read reasonably as essentially all major military powers, is pursuing them. True: credible reporting (UN reports, defence ministries, major outlets, think tanks) documents that drones with autonomous or AI-assisted targeting have been fielded or used in combat, and that most major militaries have such programmes. Mostly true: both parts hold broadly, but the autonomy is mainly AI-assisted rather than fully autonomous. Mixed: one part holds and the other does not, for example AI drones exist but only a handful of countries pursue them. Mostly false: drones exist but have essentially no AI or autonomous lethal capability, so the claim stretches remote-piloted drones into 'killer AI'. False: there is no evidence of AI-enabled lethal drones and no widespread pursuit of them.
Evidence
- When an Algorithm Controls War: How Artificial Intelligence Is Changing the Modern Battlefield E-Bezpečí · 20 April 2026
“Documented cases of use are still rare, but the first was probably in Libya in 2020, where, according to a report by the UN Security Council Panel of Experts, a Turkish Kargu-2 drone allegedly attacked a human autonomously.”
- Lethal autonomous weapon - Wikipedia Wikipedia · 10 September 2012
“As of 2025, most military drones (including unmanned aerial vehicles and unmanned combat aerial vehicles) and military robots are not truly autonomous, as they still require human control.”
- These Are the Lethal Autonomous Weapons That Terrify the U.S. Military Popular Mechanics · 3 June 2025
“The spread of decades-old precision guidance technology with advances in artificial intelligence and commercial manufacturing means that nearly every armed actor around the world now has access to rudimentary forms of AI-enabled strike capabilities.”
Reasoning
E2 cites a UN Security Council Panel of Experts report saying a Kargu-2 drone allegedly attacked a person autonomously in Libya in 2020, though it says documented cases are rare. E3 states that as of 2025 most military drones were not truly autonomous and still needed human control. E4 reports that nearly every armed actor has access to rudimentary AI-enabled strike capabilities, which broadly but indirectly supports widespread pursuit. The first part is supported by E2, which reports a UN panel finding of a probable autonomous drone attack; the second is supported only indirectly, since E4 speaks to access rather than active programmes; and E3 indicates most drones still require human control. The report judges this closest to the mostly-true criterion.
Cite this claim
Balaji Srinivasan, “Killer AI already exists in the form of drones, and every country is pursuing it.” (said 3 August 2025). Verdict: mostly true. Reliability API, claim c_2f6470b90bc5, method 0.1.0+c876030c. https://epistacklabs.com/claims/c_2f6470b90bc5
Claims without a verdict
- Unresolvable 148
- Not yet due 21
- Own records only 2
- Not researched 227
- Not checkable 1,091
Is a verdict here wrong, or do you speak for Balaji Srinivasan? Reply on the claim's own page, with the form at its foot (how replies work). Verdicts are made automatically; how verdicts and scores are made.