BRENNEN

Industries / Food, farming and forestry

Tobacco

The tobacco industry grows and cures tobacco leaf, processes it, and makes cigarettes and other tobacco and nicotine products under laws that tax, label and track them.

1612

At Jamestown, Virginia, John Rolfe planted tobacco grown from seed brought from the Caribbean, possibly Trinidad, and shipped the crop to England, where it compared well with Spanish leaf. Ralph Hamor, Secretary of Virginia, credited Rolfe with the trial, and by the National Park Service's account it became the colony's first profitable export, 20,000 pounds by 1617.

nps.gov

40

Milestones achieved across 6 domains.

10

To watch for: announced, not yet confirmed outside.

13

Open deliverables. Nobody has shipped them.

In one minute

Where it started

People in North America were using tobacco about 12,300 years ago, by the earliest evidence found, from the Wishbone site in Utah (Duke and colleagues, 2022). The industry began when John Rolfe grew leaf for sale in Virginia in 1612 and shipped it to England. Bright-leaf flue curing (1839) and James Bonsack's cigarette machine (1881) then made cheap, mass-produced cigarettes possible.

Where it stands

Farmers grew 6.05 million tonnes of tobacco leaf on 3.3 million hectares in 2024, by FAO's count, 37% of it in China, followed by India, Brazil, Indonesia and Zimbabwe. The World Health Organization (WHO) counts 1.2 billion tobacco users in 2024, down from 1.38 billion in 2000, at least 86 million adults who vape, and more than 7 million deaths a year from tobacco. A WHO treaty in force since 2005 has 183 Parties; its 2012 protocol, with 74 Parties, requires them to trace packs from factory to sale, and since 2019 every pack of cigarettes made for the EU market must carry a unique code.

What is open

Much of the industry's checking is still done by eye or by hand: classifiers grade leaf and so set the farmer's price, fieldworkers code discarded packs and pack warnings one by one, and most estimates of the illegal market come from studies the industry pays for. Open jobs include measuring from satellite images the woodland burned to cure leaf, testing how often human leaf graders agree, measuring pack warnings from photos, and an open, independent estimate of illegal cigarette sales for each EU country. Free satellite images, phone cameras and AI models that read images and documents put these within reach of a small team.

Domains

AchievedWatch forOpen

Open in tobacco

3 of 13 open deliverables

Open · Growing and curing

Curing wood counted from space

A repeatable satellite method that maps tobacco fields season by season in one growing district and estimates how much woodland is cleared each year to fuel the barns that cure them, with error bars.

Built with what exists

Free Sentinel-1 radar and Sentinel-2 optical images from the European Copernicus program; the Global Forest Change tree-loss layer from the University of Maryland; Google Earth Engine; random-forest classifiers or an open Earth-observation model such as NASA and IBM's Prithvi-EO-2.0 on Hugging Face; high-resolution basemaps for hand-labeling barns and clearings.

Credit · unclaimed

Open · Leaf processing

A grader agreement audit

A blind regrading test that measures how often trained human classifiers agree on the grade of the same leaf, published with the photos, so that grading machines can be judged against a known human range.

Built with what exists

A phone camera and a light box; a spreadsheet or a free labeling tool such as Label Studio; Fleiss' kappa, a standard measure of agreement among several raters; the Tanzanian dataset and open models in PyTorch to test machine graders against the human range.

Credit · unclaimed

Open · Cigarette making

Brand from a butt photo

A model that names the brand of a littered cigarette butt from a phone photo, using what survives on the butt: the printing and color of the tipping paper (the paper wrapped round the filter), colored bands, ventilation holes and filter type. Publish it with an open set of labeled photos, so that litter audits can count butts by maker, which producer-pays schemes need when they split cleanup costs.

Built with what exists

phone cameras with a clip-on macro lens; a labeling tool such as Label Studio or CVAT; an object detector such as Ultralytics YOLO11 to find the butt in the frame; an image classifier, or a vision-language model such as Gemini, Claude or Qwen2.5-VL to read the printing.

Credit · unclaimed