predictive AI
In plain wordsAI used to forecast outcomes about people or events, such as who will repay a loan, which patient may get sicker or which machine will fail, in order to guide decisions.
AI systems that estimate the likelihood of future or unknown outcomes from historical data in order to inform or automate decisions: credit scoring, insurance pricing, fraud detection, demand forecasting, predictive maintenance, hiring screens, criminal-justice risk assessment and clinical deterioration alerts. Technically these are mostly supervisedsupervised learning, n.: Teaching a computer by example: you give it many inputs paired with the right answers, such as photos labeled 'cat' or 'dog', and it learns to answer correctly for new inputs. discriminative modelsdiscriminative model, n.: A model that learns only to tell things apart — spam or not spam, cat or dog — without learning how to produce new examples of those things. on tabular datatabular data, n.: Data arranged as a table, like a spreadsheet: each row is a case — a customer, a loan, a patient — and each column a measured property. Much business and scientific data looks like this., often boosted treesgradient boosting, n.: A method that builds a strong predictor by adding small decision trees one at a time, each trained to correct the errors that remain after all the trees before it. or logistic regressionlogistic regression, n.: A simple, widely used method for yes-or-no predictions that outputs a probability — say, a 70% chance a customer cancels — by passing a weighted sum of inputs through an S-shaped curve.. The term is used chiefly as a contrast with generative AIgenerative AI, n.: AI that produces something new on request, such as an essay, a picture, a song or a piece of code, having absorbed from enormous numbers of existing examples what such things typically look like.. Critics, notably Arvind Narayanan and Sayash Kapoor, argue that predictive AI applied to individual life outcomes is often much less accurate than claimed, can entrench historical biasesalgorithmic bias, n.: Unfairness in software that judges or serves people: it consistently produces worse results for some groups, such as women, older people or ethnic minorities, often because it learned from skewed or incomplete records. and invites people to game it; defenders point to well-validated uses such as fraud detection and equipment-failure forecasting.
Before buying a predictive AI tool for hiring, ask how its predictions were validated on people like your applicants.
Origin
'Predictive analytics' and 'predictive modeling' were established business and statistics terms by the 2000s. 'Predictive AI' spread from about 2023 as a retronym distinguishing forecasting systems from generative AI, and was used prominently in Narayanan and Kapoor's book AI Snake Oil (2024).
Usage note
Not a technical category: the same model families power both kinds of AI. In EU law many predictive uses, such as creditworthiness and employment decisions, are high-riskhigh-risk AI system, n.: Under EU law, an AI system used where mistakes could seriously hurt people's safety or rights — in hiring, schooling, loans, policing, borders or products like medical devices — and which must therefore pass strict checks., and predicting a person's risk of committing a crime solely from profiling is a prohibited practiceprohibited AI practices, n.: Uses of AI that the EU bans outright as too harmful — such as manipulating people, social scoring, and most live facial recognition by police in public places..
Related
- generative AIgenerative AI, n.: AI that produces something new on request, such as an essay, a picture, a song or a piece of code, having absorbed from enormous numbers of existing examples what such things typically look like.
- discriminative modeldiscriminative model, n.: A model that learns only to tell things apart — spam or not spam, cat or dog — without learning how to produce new examples of those things.
- predictive policingpredictive policing, n.: Using computer programs to forecast where crimes are likely to happen, or who is likely to be involved, so that police can decide where to patrol or whom to watch.
- automated decision-makingautomated decision-making, n.: When a computer system, rather than a person, makes or effectively makes a decision about someone — such as approving credit, screening job applicants or flagging possible welfare fraud.
- algorithmic biasalgorithmic bias, n.: Unfairness in software that judges or serves people: it consistently produces worse results for some groups, such as women, older people or ethnic minorities, often because it learned from skewed or incomplete records.
- time series forecastingtime series forecasting, n.: Predicting future values of something measured over time, such as sales, electricity demand, prices or rainfall, from its past values and related information.
- COMPASCOMPAS, prop. n.: Commercial software used in parts of the US justice system to rate how likely someone is to commit another crime; a 2016 news investigation asking whether it was unfair to Black defendants made it famous.
- generative AIgenerative AI, n.: AI that produces something new on request, such as an essay, a picture, a song or a piece of code, having absorbed from enormous numbers of existing examples what such things typically look like.
Sources
- Narayanan, A. & Kapoor, S. (2024). AI Snake Oil: What Artificial Intelligence Can Do, What It Can't, and How to Tell the Difference. Princeton University Press.
Mentioned in
generative AIgenerative AI, n.: AI that produces something new on request, such as an essay, a picture, a song or a piece of code, having absorbed from enormous numbers of existing examples what such things typically look like., predictionprediction, n.: A model's output for a given input: its best guess at an unknown answer. In AI it need not be about the future; labeling a photo 'cat' counts as a prediction., Arvind NarayananArvind Narayanan, prop. n.: An Indian-American computer scientist at Princeton who studies AI's real-world effects and argues that AI is a 'normal technology' whose impact will unfold gradually.