Given the increasing importance of data in today's economy, is it now time to tax data harvesting activities?
Data is arguably the ‘oil’ of today's knowledge economy and its value skyrocketed with the dawn of generative AI. It is the lifeblood for AI systems like the GPT series, which require vast quantities of internet-scraped data for development and training. This immense reliance on the public’s freely accessed data brings forth a crucial question: should not then the public be compensated for their invaluable commercial contribution? If so, there is one intriguing solution: tax the data itself.
As highlighted by the OECD’s BEPS Pillar One project, current tax systems struggle to conceptualise a fair manner in taxing the digital economy and its major players. This is because, in addition to having ethereal global footprints, a lot of their value is generated by non-monetary activities and synergies. Purchasing databases or licensing digital advertising may be easily taxable events, but the ultimate worth comes from the intangible insights and synergies obtained from the data collected, especially freely given user data. This latent value remains mostly untaxed and only really materialises when realized through the sale of company shares on capital gains.
That said, a tax on data presents challenges. Assigning a tangible monetary value to data is not straightforward. Raw data on its own is worthless without refinement. Moreover, data’s mass relational relevance and recency affect its value, which can appreciate or depreciate over time.
Nevertheless, data tax proposals do exist. New York's Bill 4959 suggested in 2021 a tiered, per-head monthly excise tax on data collectors, starting at five cents/month per NY-based consumer. Massachusetts legislators, similarly, in early 2023, introduced a series of bills targeting the digital economy for taxation. These proposals ranged from implementing a 6.25% excise tax on digital advertising revenue with a $1m exemption, to a tiered tax on data collection, exempting the first $1m in revenue but increasing rates based on the number of consumers.
Tax legal experts have also weighed in with various models. Legal scholar Reuven Avi-Yonah champions a data excise tax based on volume of data downloaded, rather than the NY per head count model. While Facebook co-founder Chris Hughes proposes a ‘Data Dividend’ model, taxing significant data collectors and then redistributing the proceeds to the public much like Alaska’s Permanent Fund does with the extractives industry.
In a similar spirit, University of California Professor Omri Maran's royalty-based strategy may be preferred over taxes entirely. This would be likening data to a commodity in the extractive industry. Such a system would offer a flexible and indefinitely renewable revenue stream.
Whatever form a data tax takes, it is imperative that they complement in tandem current privacy and intellectual property rights initiatives. The insatiable demand for data by generative AI raises critical equity concerns, particularly when the data in question is generated by the public. This requires a multifaceted approach regarding regulation. The GDPR and the EU’s AI Act, while bolstering digital security and protections against misuse, are only parts of an equitable solution. Considering only data security misses the tremendous monetary worth that data will have in the age of generative AI. These legal frameworks on their own fail to compensate individuals for their significant contributions to this new digital frontier.
Moreover, an international consensus is essential.Current international tax blueprints, like Pillar One, scarcely mention artificial intelligence, making their proposals already feel outdated. Because data is simultaneously omnipresent and intangible, global collaboration is paramount. Countries must ensure that as technology evolves, international tax frameworks are not left behind.
Data’s evolving role in powering such a technological leap as generative AI necessitates an evolution in our taxation policies. This relationship will be crucial in the 21st century. Tax solutions must be flexible, recognising the worth of this abundant and increasingly vital technological resource.
Matthew Chin Barnes, Linklaters’ Caribbean tax internship programme
Given the increasing importance of data in today's economy, is it now time to tax data harvesting activities?
Data is arguably the ‘oil’ of today's knowledge economy and its value skyrocketed with the dawn of generative AI. It is the lifeblood for AI systems like the GPT series, which require vast quantities of internet-scraped data for development and training. This immense reliance on the public’s freely accessed data brings forth a crucial question: should not then the public be compensated for their invaluable commercial contribution? If so, there is one intriguing solution: tax the data itself.
As highlighted by the OECD’s BEPS Pillar One project, current tax systems struggle to conceptualise a fair manner in taxing the digital economy and its major players. This is because, in addition to having ethereal global footprints, a lot of their value is generated by non-monetary activities and synergies. Purchasing databases or licensing digital advertising may be easily taxable events, but the ultimate worth comes from the intangible insights and synergies obtained from the data collected, especially freely given user data. This latent value remains mostly untaxed and only really materialises when realized through the sale of company shares on capital gains.
That said, a tax on data presents challenges. Assigning a tangible monetary value to data is not straightforward. Raw data on its own is worthless without refinement. Moreover, data’s mass relational relevance and recency affect its value, which can appreciate or depreciate over time.
Nevertheless, data tax proposals do exist. New York's Bill 4959 suggested in 2021 a tiered, per-head monthly excise tax on data collectors, starting at five cents/month per NY-based consumer. Massachusetts legislators, similarly, in early 2023, introduced a series of bills targeting the digital economy for taxation. These proposals ranged from implementing a 6.25% excise tax on digital advertising revenue with a $1m exemption, to a tiered tax on data collection, exempting the first $1m in revenue but increasing rates based on the number of consumers.
Tax legal experts have also weighed in with various models. Legal scholar Reuven Avi-Yonah champions a data excise tax based on volume of data downloaded, rather than the NY per head count model. While Facebook co-founder Chris Hughes proposes a ‘Data Dividend’ model, taxing significant data collectors and then redistributing the proceeds to the public much like Alaska’s Permanent Fund does with the extractives industry.
In a similar spirit, University of California Professor Omri Maran's royalty-based strategy may be preferred over taxes entirely. This would be likening data to a commodity in the extractive industry. Such a system would offer a flexible and indefinitely renewable revenue stream.
Whatever form a data tax takes, it is imperative that they complement in tandem current privacy and intellectual property rights initiatives. The insatiable demand for data by generative AI raises critical equity concerns, particularly when the data in question is generated by the public. This requires a multifaceted approach regarding regulation. The GDPR and the EU’s AI Act, while bolstering digital security and protections against misuse, are only parts of an equitable solution. Considering only data security misses the tremendous monetary worth that data will have in the age of generative AI. These legal frameworks on their own fail to compensate individuals for their significant contributions to this new digital frontier.
Moreover, an international consensus is essential.Current international tax blueprints, like Pillar One, scarcely mention artificial intelligence, making their proposals already feel outdated. Because data is simultaneously omnipresent and intangible, global collaboration is paramount. Countries must ensure that as technology evolves, international tax frameworks are not left behind.
Data’s evolving role in powering such a technological leap as generative AI necessitates an evolution in our taxation policies. This relationship will be crucial in the 21st century. Tax solutions must be flexible, recognising the worth of this abundant and increasingly vital technological resource.
Matthew Chin Barnes, Linklaters’ Caribbean tax internship programme
