From d60e1a42ecf4ac3fa71461bc4ad7371286a7cea8 Mon Sep 17 00:00:00 2001 From: Edouard Mathieu Date: Mon, 25 Nov 2024 16:12:03 +0100 Subject: [PATCH] =?UTF-8?q?=E2=9C=A8=20=20improve=20metadata=20for=20AI=20?= =?UTF-8?q?investment?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .../2024-07-16/cset.meta.yml | 20 +++++++++---------- .../ai_corporate_investment.meta.yml | 20 +++++++++---------- ...i_investment_generative_companies.meta.yml | 16 +++++++-------- .../2024-06-28/ai_private_investment.meta.yml | 18 ++++++++--------- 4 files changed, 35 insertions(+), 39 deletions(-) diff --git a/etl/steps/data/garden/artificial_intelligence/2024-07-16/cset.meta.yml b/etl/steps/data/garden/artificial_intelligence/2024-07-16/cset.meta.yml index 7cd567eab30..2faebe8865c 100644 --- a/etl/steps/data/garden/artificial_intelligence/2024-07-16/cset.meta.yml +++ b/etl/steps/data/garden/artificial_intelligence/2024-07-16/cset.meta.yml @@ -9,13 +9,13 @@ definitions: numDecimalPlaces: 0 description_processing_investment: |- - - Reporting a time series of AI investments in nominal prices (i.e., without adjusting for inflation) means it makes little sense to compare observations across time; it is therefore not very useful. To make comparisons across time possible, one has to take into account that prices change (e.g., there is inflation). - - It is not obvious how to adjust this time series for inflation, and we debated it at some length within our team. - - It would be straightforward to adjust the time series for price changes if we knew the prices of the specific goods and services that these investments purchased. This would make it possible to calculate a volume measure of AI investments, and it would tell us how much these investments bought. But such a metric is not available. While a comprehensive price index is not available, we know that the cost for some crucial AI technology has fallen rapidly in price. - - In the absence of a comprehensive price index that captures the price of AI-specific goods and services, one has to rely on one of the available metrics for the price of a bundle of goods and services. In the end we decided to use the US Consumer Price Index (CPI). - - The US CPI does not provide us with a volume measure of AI goods and services, but it does capture the opportunity costs of these investments. The inflation adjustment of this time series of AI investments therefore lets us understand the size of these investments relative to whatever else these sums of money could have purchased. + - Reporting a time series of AI investments in nominal prices would make it difficult to compare observations across time. To make these comparisons possible, one has to take into account that prices change (inflation). + - It is not obvious how to adjust this time series for inflation, and our team discussed the best solutions at our disposal. + - It would be straightforward to adjust the time series for price changes if we knew the prices of the specific goods and services purchased through these investments. This would make it possible to calculate a volume measure of AI investments and tell us how much these investments bought. But such a metric is not available. While a comprehensive price index is not available, we know that the cost of some crucial AI technology has fallen rapidly in price. + - In the absence of a comprehensive price index that captures the price of AI-specific goods and services, one has to rely on one of the available metrics for the price of a bundle of goods and services. Ultimately, we decided to use the US Consumer Price Index (CPI). + - The US CPI does not provide us with a volume measure of AI goods and services, but it does capture the opportunity costs of these investments. The inflation adjustment of this time series of AI investments, therefore, lets us understand the size of these investments relative to whatever else these sums of money could have purchased. - description_short_investment: Only includes private-market investment flows, such as venture capital; excludes all investment in publicly traded companies, such as the "Big Tech" firms. This data is expressed in US dollars, adjusted for inflation. + description_short_investment: Only includes private-market investment such as venture capital; excludes all investment in publicly traded companies, such as "Big Tech" firms. This data is expressed in US dollars, adjusted for inflation. description_short_articles: English- and Chinese-language scholarly publications related to the development and application of AI. This includes journal articles, conference papers, repository publications (such as arXiv), books, and theses. @@ -35,12 +35,12 @@ definitions: World aggregate does not include data for Micronesia, Tonga, Samoa, Kiribati, Fiji, Papua New Guinea, Palau, Tuvalu, Bermuda, Armenia, Belarus, Georgia, Gibraltar, Jersey, Kosovo, Moldova, Isle of Man, Iceland, Albania, Andorra, Bosnia and Herzegovina, Malta, Montenegro, San Marino, North Macedonia, Liechtenstein, Monaco, Vatican City, Guernsey, Afghanistan, Kyrgyzstan, Bahrain, Laos, Bangladesh, Lebanon, Bhutan, Maldives, Cambodia, Syria, Tajikistan, Cyprus, Mongolia, North Korea, Myanmar, Timor-Leste, Nepal, Turkmenistan, Pakistan, Palestine, Iraq, United Arab Emirates, Uzbekistan, Kazakhstan, Qatar, Vietnam, Yemen, Kuwait, Algeria, Cape Verde, Equatorial Guinea, Swaziland, Namibia, Central African Republic (the), Angola, Ethiopia, Niger, Benin, Gabon, Nigeria, Botswana, Gambia, Rwanda, Burkina Faso, Ghana, São Tomé and Príncipe, Burundi, Guinea, Senegal, Guinea-Bissau, Seychelles, Cameroon, Sierra Leone, Lesotho, Somalia, Chad, Liberia, Libya, South Sudan, Congo, Madagascar, Sudan, Côte d'Ivoire, Malawi, Togo, Mali, Djibouti, Mauritania, Uganda, Egypt, Mauritius, Tanzania, Zambia, Eritrea, Mozambique, Zimbabwe, Comoros, Antigua and Barbuda, Bolivia, Suriname, Nicaragua, Dominican Republic, Bahamas, Ecuador, Paraguay, Barbados, Saint Vincent and the Grenadines, El Salvador, Belize, Grenada, Saint Kitts and Nevis, Guatemala, Guyana, Haiti, Honduras, Trinidad and Tobago, Jamaica, Venezuela, Puerto Rico, Cayman Islands (the), Turks and Caicos Islands, Saint Lucia, and Dominica. description_key_investment: &description_key_investment |- - - The dataset only covers private-market investment flows, such as venture capital. It excludes non-equity financing, such as debt and grants, and omits publicly traded companies, including major Big Tech firms (e.g., Amazon, Microsoft, Meta). As a result, significant investments from public companies, corporate R&D, government funding, and broader infrastructure costs (like data centers and hardware) are not captured, limiting the dataset’s coverage of global AI investments. + - The data likely underestimates total global AI investment, as it only captures certain types of private equity transactions, excluding other significant channels and categories of AI-related spending. + - The dataset only covers private-market investment such as venture capital. It excludes non-equity financing, such as debt and grants, and publicly traded companies, including major Big Tech firms. As a result, significant investments from public companies, corporate R&D, government funding, and broader infrastructure costs (like data centers and hardware) are not captured, limiting the data's coverage of global AI investments. + - The data's "World" aggregate reflects the total investment represented in the data, but may not represent global AI efforts comprehensively, especially in countries not included in the data. - Companies are classified as AI-related based on keyword and industry tags, potentially including firms not traditionally seen as AI-focused while missing others due to definitional differences. - Many investment values are undisclosed, so the source relies on median values from similar transactions, introducing some uncertainty. Additionally, investment origin is attributed to company headquarters, which may overlook cross-border structures or varied investor origins. - - One-time events like large acquisitions can skew yearly figures, and macroeconomic conditions (e.g., interest rates, market sentiment) may impact trends independently of AI-related dynamics. - - The dataset’s "World" aggregate reflects the total investment represented but does not encompass global AI efforts comprehensively, especially in countries not included in the data. - - The dataset likely underestimates the total global AI investment, as it only captures certain types of private equity transactions, excluding other significant channels and categories of AI-related spending. + - One-time events, such as large acquisitions, can distort yearly figures, while broader economic factors like interest rates and market sentiment can influence investment trends independently of AI-specific developments. # Learn more about the available fields: # http://docs.owid.io/projects/etl/architecture/metadata/reference/ diff --git a/etl/steps/data/grapher/artificial_intelligence/2024-06-28/ai_corporate_investment.meta.yml b/etl/steps/data/grapher/artificial_intelligence/2024-06-28/ai_corporate_investment.meta.yml index d1491908586..cf5ceb40082 100644 --- a/etl/steps/data/grapher/artificial_intelligence/2024-06-28/ai_corporate_investment.meta.yml +++ b/etl/steps/data/grapher/artificial_intelligence/2024-06-28/ai_corporate_investment.meta.yml @@ -9,19 +9,19 @@ definitions: note: This data is expressed in constant 2021 US$. Inflation adjustment is based on the US Consumer Price Index (CPI). description_processing: |- - - Reporting a time series of AI investments in nominal prices (i.e., without adjusting for inflation) means it makes little sense to compare observations across time; it is therefore not very useful. To make comparisons across time possible, one has to take into account that prices change (e.g., there is inflation). - - It is not obvious how to adjust this time series for inflation, and we debated it at some length within our team. - - It would be straightforward to adjust the time series for price changes if we knew the prices of the specific goods and services that these investments purchased. This would make it possible to calculate a volume measure of AI investments, and it would tell us how much these investments bought. But such a metric is not available. While a comprehensive price index is not available, we know that the cost for some crucial AI technology has fallen rapidly in price. - - In the absence of a comprehensive price index that captures the price of AI-specific goods and services, one has to rely on one of the available metrics for the price of a bundle of goods and services. In the end we decided to use the US Consumer Price Index (CPI). - - The US CPI does not provide us with a volume measure of AI goods and services, but it does capture the opportunity costs of these investments. The inflation adjustment of this time series of AI investments therefore lets us understand the size of these investments relative to whatever else these sums of money could have purchased. + - Reporting a time series of AI investments in nominal prices would make it difficult to compare observations across time. To make these comparisons possible, one has to take into account that prices change (inflation). + - It is not obvious how to adjust this time series for inflation, and our team discussed the best solutions at our disposal. + - It would be straightforward to adjust the time series for price changes if we knew the prices of the specific goods and services purchased through these investments. This would make it possible to calculate a volume measure of AI investments and tell us how much these investments bought. But such a metric is not available. While a comprehensive price index is not available, we know that the cost of some crucial AI technology has fallen rapidly in price. + - In the absence of a comprehensive price index that captures the price of AI-specific goods and services, one has to rely on one of the available metrics for the price of a bundle of goods and services. Ultimately, we decided to use the US Consumer Price Index (CPI). + - The US CPI does not provide us with a volume measure of AI goods and services, but it does capture the opportunity costs of these investments. The inflation adjustment of this time series of AI investments, therefore, lets us understand the size of these investments relative to whatever else these sums of money could have purchased. description_key: + - The data likely underestimates total global AI investment, as it only captures certain types of private equity transactions, excluding other significant channels and categories of AI-related spending. + - This data focuses on traditional corporate finance deals, but the source does not fully disclose its methodology and what's included or excluded. This means it may not fully capture important areas of AI investment, such as those from publicly traded companies, corporate internal R&D, government funding, public sector initiatives, data center infrastructure, hardware production, semiconductor manufacturing, and expenses for research and talent. + - One-time events, such as large acquisitions, can distort yearly figures, while broader economic factors like interest rates and market sentiment can influence investment trends independently of AI-specific developments. - A merger is a corporate strategy involving two companies joining together to form a new company. An acquisition is a corporate strategy involving one company buying another company. - - Private investment in AI companies in each year that received an investment of more than $1.5 million (not adjusted for inflation). + - Private investment is defined as investment in AI companies of more than $1.5 million (in current US dollars). - A public offering is the sale of equity shares or other financial instruments to the public in order to raise capital. - A minority stake is an ownership interest of less than 50% of the total shares of a company. - - The categories shown suggest a focus on traditional corporate finance deals, but without a detailed methodology, we can't be certain about what's included or excluded. This means it may not fully capture important areas of AI investment, such as those from public companies (e.g., NVIDIA, TSMC), corporate internal R&D, government funding, public sector initiatives, data center infrastructure, hardware production, semiconductor manufacturing, and expenses for research and talent. - - One-time events like large acquisitions can skew yearly figures, and broader economic factors like interest rates or market sentiment can also affect AI investment trends independently of AI-specific developments. - - The dataset likely underestimates the total global AI investment, as it only captures certain types of private equity transactions, excluding other significant channels and categories of AI-related spending. description_short: This data is expressed in US dollars, adjusted for inflation. unit: 'constant 2021 US$' @@ -38,5 +38,3 @@ tables: variables: world: title: Global corporate investment in AI - - diff --git a/etl/steps/data/grapher/artificial_intelligence/2024-06-28/ai_investment_generative_companies.meta.yml b/etl/steps/data/grapher/artificial_intelligence/2024-06-28/ai_investment_generative_companies.meta.yml index d6866c816d2..5ce51b53dd1 100644 --- a/etl/steps/data/grapher/artificial_intelligence/2024-06-28/ai_investment_generative_companies.meta.yml +++ b/etl/steps/data/grapher/artificial_intelligence/2024-06-28/ai_investment_generative_companies.meta.yml @@ -21,15 +21,15 @@ tables: short_unit: '$' description_short: Generative AI refers to AI systems that can create new content, such as images, text, or music, based on patterns learned from existing data. description_processing: |- - - Reporting a time series of AI investments in nominal prices (i.e., without adjusting for inflation) means it makes little sense to compare observations across time; it is therefore not very useful. To make comparisons across time possible, one has to take into account that prices change (e.g., there is inflation). - - It is not obvious how to adjust this time series for inflation, and we debated it at some length within our team. - - It would be straightforward to adjust the time series for price changes if we knew the prices of the specific goods and services that these investments purchased. This would make it possible to calculate a volume measure of AI investments, and it would tell us how much these investments bought. But such a metric is not available. While a comprehensive price index is not available, we know that the cost for some crucial AI technology has fallen rapidly in price. - - In the absence of a comprehensive price index that captures the price of AI-specific goods and services, one has to rely on one of the available metrics for the price of a bundle of goods and services. In the end we decided to use the US Consumer Price Index (CPI). - - The US CPI does not provide us with a volume measure of AI goods and services, but it does capture the opportunity costs of these investments. The inflation adjustment of this time series of AI investments therefore lets us understand the size of these investments relative to whatever else these sums of money could have purchased. + - Reporting a time series of AI investments in nominal prices would make it difficult to compare observations across time. To make these comparisons possible, one has to take into account that prices change (inflation). + - It is not obvious how to adjust this time series for inflation, and our team discussed the best solutions at our disposal. + - It would be straightforward to adjust the time series for price changes if we knew the prices of the specific goods and services purchased through these investments. This would make it possible to calculate a volume measure of AI investments and tell us how much these investments bought. But such a metric is not available. While a comprehensive price index is not available, we know that the cost of some crucial AI technology has fallen rapidly in price. + - In the absence of a comprehensive price index that captures the price of AI-specific goods and services, one has to rely on one of the available metrics for the price of a bundle of goods and services. Ultimately, we decided to use the US Consumer Price Index (CPI). + - The US CPI does not provide us with a volume measure of AI goods and services, but it does capture the opportunity costs of these investments. The inflation adjustment of this time series of AI investments, therefore, lets us understand the size of these investments relative to whatever else these sums of money could have purchased. description_key: - - One-time events like large acquisitions can skew yearly figures, and broader economic factors like interest rates or market sentiment can also affect AI investment trends independently of AI-specific developments. - - The dataset’s methodology doesn’t specify which types of AI investments are included, so it may overlook important areas of AI investment, such as those from public companies (e.g., NVIDIA, TSMC), corporate internal R&D, government funding, public sector initiatives, data center infrastructure, hardware production, semiconductor manufacturing, and expenses for research and talent. - - The dataset likely underestimates the total global AI investment, as it only captures certain types of private equity transactions, excluding other significant channels and categories of AI-related spending. + - The data likely underestimates total global AI investment, as it only captures certain types of private equity transactions, excluding other significant channels and categories of AI-related spending. + - The source does not fully disclose its methodology and what's included or excluded. This means it may not fully capture important areas of AI investment, such as those from publicly traded companies, corporate internal R&D, government funding, public sector initiatives, data center infrastructure, hardware production, semiconductor manufacturing, and expenses for research and talent. + - One-time events, such as large acquisitions, can distort yearly figures, while broader economic factors like interest rates and market sentiment can influence investment trends independently of AI-specific developments. presentation: grapher_config: diff --git a/etl/steps/data/grapher/artificial_intelligence/2024-06-28/ai_private_investment.meta.yml b/etl/steps/data/grapher/artificial_intelligence/2024-06-28/ai_private_investment.meta.yml index 717ac9b8266..e99162560b3 100644 --- a/etl/steps/data/grapher/artificial_intelligence/2024-06-28/ai_private_investment.meta.yml +++ b/etl/steps/data/grapher/artificial_intelligence/2024-06-28/ai_private_investment.meta.yml @@ -9,17 +9,17 @@ definitions: note: This data is expressed in constant 2021 US$. Inflation adjustment is based on the US Consumer Price Index (CPI). description_processing: |- - - Reporting a time series of AI investments in nominal prices (i.e., without adjusting for inflation) means it makes little sense to compare observations across time; it is therefore not very useful. To make comparisons across time possible, one has to take into account that prices change (e.g., there is inflation). - - It is not obvious how to adjust this time series for inflation, and we debated it at some length within our team. - - It would be straightforward to adjust the time series for price changes if we knew the prices of the specific goods and services that these investments purchased. This would make it possible to calculate a volume measure of AI investments, and it would tell us how much these investments bought. But such a metric is not available. While a comprehensive price index is not available, we know that the cost for some crucial AI technology has fallen rapidly in price. - - In the absence of a comprehensive price index that captures the price of AI-specific goods and services, one has to rely on one of the available metrics for the price of a bundle of goods and services. In the end we decided to use the US Consumer Price Index (CPI). - - The US CPI does not provide us with a volume measure of AI goods and services, but it does capture the opportunity costs of these investments. The inflation adjustment of this time series of AI investments therefore lets us understand the size of these investments relative to whatever else these sums of money could have purchased. + - Reporting a time series of AI investments in nominal prices would make it difficult to compare observations across time. To make these comparisons possible, one has to take into account that prices change (inflation). + - It is not obvious how to adjust this time series for inflation, and our team discussed the best solutions at our disposal. + - It would be straightforward to adjust the time series for price changes if we knew the prices of the specific goods and services purchased through these investments. This would make it possible to calculate a volume measure of AI investments and tell us how much these investments bought. But such a metric is not available. While a comprehensive price index is not available, we know that the cost of some crucial AI technology has fallen rapidly in price. + - In the absence of a comprehensive price index that captures the price of AI-specific goods and services, one has to rely on one of the available metrics for the price of a bundle of goods and services. Ultimately, we decided to use the US Consumer Price Index (CPI). + - The US CPI does not provide us with a volume measure of AI goods and services, but it does capture the opportunity costs of these investments. The inflation adjustment of this time series of AI investments, therefore, lets us understand the size of these investments relative to whatever else these sums of money could have purchased. description_short: Includes companies that received more than $1.5 million in investment (not adjusted for inflation). This data is expressed in US dollars, adjusted for inflation. description_key: - - One-time events like large acquisitions can skew yearly figures, and broader economic factors like interest rates or market sentiment can also affect AI investment trends independently of AI-specific developments. - - The dataset’s methodology doesn’t specify which types of AI investments are included, so it may overlook important areas of AI investment, such as those from public companies (e.g., NVIDIA, TSMC), corporate internal R&D, government funding, public sector initiatives, data center infrastructure, hardware production, semiconductor manufacturing, and expenses for research and talent. - - The dataset likely underestimates the total global AI investment, as it only captures certain types of private equity transactions, excluding other significant channels and categories of AI-related spending. + - The data likely underestimates total global AI investment, as it only captures certain types of private equity transactions, excluding other significant channels and categories of AI-related spending. + - The source does not fully disclose its methodology and what's included or excluded. This means it may not fully capture important areas of AI investment, such as those from publicly traded companies, corporate internal R&D, government funding, public sector initiatives, data center infrastructure, hardware production, semiconductor manufacturing, and expenses for research and talent. + - One-time events, such as large acquisitions, can distort yearly figures, while broader economic factors like interest rates and market sentiment can influence investment trends independently of AI-specific developments. unit: 'constant 2021 US$' short_unit: '$' @@ -41,5 +41,3 @@ tables: title: Private investment in AI in the United States european_union_and_united_kingdom: title: Private investment in AI in the European Union and United Kingdom - -