On Floods, disaster management & AI Dividend

Just 2 links (+ a bonus link) this weekend but hopefully all interesting and thought-provoking.

The first is a sharp & insightful piece by Shankkar Aiyar on how systemic inertia, inaction and lack of accountability lead to an event becoming a disaster. Below is a brief catalogue of some sobering facts from the article in the context of the disastrous floods in Delhi and other parts of India:

….A report by Asian Development Bank states “India experienced 278 floods during 1980–2017 that affected more than 750 million people and caused about $58.7 billion in losses.” As per NCRB records between 2011 and 2021, disasters accounted for the death of 145,514 persons –that is, roughly 36 persons lost their lives every day in the period in what the government terms as deaths “caused by nature”. If it wasn’t for the dedication of those in uniform – from the armed forces and the NDRF – the toll would be decidedly higher.

…India hosts 34,919 glaciers and 6921 glacial lakes in the Himalayan Region. How are they managed? From inventory to research to forecasting there are nine departments and ministries involved. The recommendation to bring them under one agency – given the acceleration in global warming and risks – is pending. Meanwhile the annual rate of retreat of glaciers is estimated between 5 and 20 metres. The Himalayan-Karakoram region is warming at a faster rate than the global mean by 0.5 degree C. This will impact local changes in hydrology, in terms of rainfall, snowfall patterns and increased melting from glaciers trigger a rise in extreme events leading to disasters.

…. the Standing Committee of Parliament on Water Resources reveals only three states have adopted draft legislation for flood plains zoning while the states most at risk have ignored recommendations. Soil erosion worsens weather events but states such as Assam, home to the mighty Brahmaputra, have made no effort to tackle it. Worse allocations for flood management have been slashed to a third from Rs 14,500 crore.

…How widely are the early warning systems deployed? A CEEW Study shows that approximately two of three persons or 66 per cent of the population in India is exposed to extreme flood events but only 33 per cent of the exposed populace is covered by flood early warning systems. Critically India lacks multi-hazard warning systems. The Met Department which forecasts heavy rainfall and cloudbursts does not have early warning systems for glacial lake outburst, floods, landslides, avalanches etc

Deluge of Rhetoric Drowns Accountability- The New Indian Express

The second link is provocative article by Barath Raghavan & Bruce Schneier on how we can claim a share of the profits for all the data that powers today’s AI and related technologies, including ChatGPT etc.

This has been proposed before but I have not come across any instance where someone has actually got paid for their content (including writing, comments, tweets, images). This is not as trivial as you might think. Recall that Musk recently threatened Microsoft over claims that the company “trained illegally using Twitter data” and has recently filed a case against 4 individuals it accuses of scraping twitter data.

So how exactly can we get Big Tech to pay for all the data it used for free to train its large AI models? Bruce and Barath propose an AI Dividend, based on the lines of Alaska’s Permanent Fund that pays Alaskans a share of the profits derived from oil in their state. Interesting idea but I wonder whether it will get any traction. Regardless, it is worthy of a wider discussion.

The bonus link is this BBC report on Skywells in old Chinese homes to keep them cool during summer. Apparently this is not limited to China. Ancient Romans had been toying with it too! I have not come across anything similar in traditional Indian architecture, although the courtyard in the middle of traditional homes probably served a similar purpose.