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?
74 68 79 -
Predictions
Statements about events outside the person's control, checked once they fell due: did they come true?
Insufficient record: 1 prediction resolved, 5 needed for a score.
-
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: 74 (likely 68 to 79, high confidence), from 97 claims: 97 factual claims. It mixes kinds of claim in different proportions for different people, so compare people kind by kind.
Steven Pinker's claims
15 claims
“Part of the problem is that a lot of artificial intelligence has been deployed toward dubious goals like multiplying propaganda, disinformation, phishing scams, manipulating people into doing things against their interests.”
Steven Pinker, interviewee 2023 The Intelligence and Rationality of AI and Humans: A Conversation With Steven Pinker (Harvard Data Science Review)
A lot of artificial intelligence has been deployed toward dubious goals such as propaganda, disinformation, phishing scams and manipulating people.
Mixed Partly true and partly false in roughly equal measure.
What would count as true or false
Judge as of 2023. True: credible documentation from that year, such as security-industry reports, platform threat reports, government or NGO reports (for example Freedom House 2023) or academic studies, shows AI being used widely (not in isolated cases) for propaganda or disinformation, phishing or scams, and manipulative targeting. Mostly true: widespread use is documented for most of the listed categories, but one is thinly documented. Mixed: the uses are documented but only as scattered cases or emerging trends, so 'a lot' is only partly supported. Mostly false: such uses were mainly hypothetical or anticipated in 2023, with only a few real instances. False: there is no documented deployment of AI for these purposes.
Evidence
- FREEDOM ON THE NET 2023 Freedom House · 2023
“During the February 2023 Nigerian elections, an AI-manipulated audio clip spread on social media, purportedly implicating an opposition presidential candidate in plans to rig balloting.”
- FREEDOM ON THE NET 2023 Freedom House · 2023
“However, much of the work is still done by humans.”
- Twice as many cyber incidents reported and rise in AI scam attempts Swiss National Cyber Security Centre (NCSC), federal government · 6 May 2024
“Although the number of reports of such incidents is still comparatively low, the NCSC believes that these are the first attempts by cyber criminals to explore how AI might be used for future cyberattacks.”
- Meta Quarterly Adversarial Threat Report Q3 2023 Meta · November 2023
“we’ve seen a case when a political party in Poland reportedly used AI-created audio in the course of an election campaign to make an argument against their opponents.”
Reasoning
E1 documents a 2023 AI-manipulated audio clip in Nigeria's election, and E4 reports a case in which a Polish party reportedly used AI-created audio in an election campaign. However, E2 notes that much of this work was still done by humans. E3 calls AI scam reports "comparatively low" and describes them as criminals' first attempts. The evidence shows scattered cases and emerging trends rather than widespread use in every listed category, and none of the evidence items addresses manipulative targeting, so "a lot" is only partly supported.
Cite this claim
Steven Pinker, “A lot of artificial intelligence has been deployed toward dubious goals such as propaganda, disinformation, phishing scams and manipulating people.” (said 2023). Verdict: mixed. Reliability API, claim c_e4b4bb430c88, method 0.1.0+28da144b. https://epistacklabs.com/claims/c_e4b4bb430c88
Claims without a verdict
Is a verdict here wrong, or do you speak for Steven Pinker? 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.