Artificial Intelligence Juggernauts and the
Data Economy that Feeds Them
The first in our series exploring the digitalization of nearly everything, this article highlights promising career directions accompanying this digital transformation that are particularly well-served by a trained lawyer’s mindset. We’ll survey the accelerating emergence of artificial intelligence and implications of a growing “data economy”. In subsequent articles, we’ll delve deeper into the alternative career pathways highlighted.
What was the first “data”? Was it some primordial occasion of “evidence” introduced in the course of an ancient ritual foreshadowing the justice systems of today? Was it some specific number of pebbles placed carefully in view between parties to the sharing of food or other essentials – marking and mediating social reciprocity among our earliest ancestors? Some would point to the innovation of the first “receipts” arguably from as far back as 7500-3500 BCE, used to secure the trade of various commodities at the dawn of human civilization. These first tokens processed data in direct, concrete terms – counting material goods and quantifying them with markings in one-to-one correspondence with the number of goods. From this early Neolithic data science of tactile token objects to their subsequent abstraction into written numerals and the evolution of modern algebra, we arrive now at an age wherein we conceive of data in terms of the extremely abstract high-dimensional vector spaces of modern-day AI systems. Whatever its origins and era of abstraction, it is undeniable that the path and character of human cultural evolution is deeply influenced by an ever-expanding universe of data and the ever more complex and powerful uses to which it’s applied. We are at the dawn of new cultures and societies in which every person, place or thing emanates a fountain of data which is stored, processed, traded, transformed and employed to train remarkably capable artificial intelligences. New unexpected challenges are emerging with this revolution in data-driven computational cognitive capacity. The “grandfathers” of these innovations are now warning that we must not underestimate their potential for harm – even speaking in terms of, “existential risk”. These kinds of tectonic shifts – affecting all industries and sectors – increasingly demand a workforce broadly equipped with commensurate ethical imagination and capacities of careful analysis as developed in the course of our legal education and career experience.
For Lawyers and grads exploring the wide range of new careers possibilities emerging with these technologies, it’s worth noting that these tools are experiencing a more rapid adoption than any previous technological innovation. In 2023, Reuters and other major international news outlets reported on a USB Bank analysis, stating “ChatGPT, the popular chatbot from OpenAI, is estimated to have reached 100 million monthly active users in January, just two months after launch, making it the fastest-growing consumer application in history, according to a UBS study on Wednesday” [2]
According to a 2022 European Commission press release, “The volume of data is constantly growing, 33 zettabytes generated in 2018 to 175 zettabytes expected in 2025. It is an untapped potential, 80% of industrial data is never used. The Data Act addresses the legal, economic and technical issues that lead to data being under-used.” Further stating that, “new rules will make more data available for reuse and are expected to create €270 billion of additional GDP by 2028.” [3]
This new economics of “data” and the synthetic intelligences so enabled intensifies the shift toward rent capture on intangible assets and creates complex new markets for information and insight. Those markets are attracting greater regulatory and compliance scrutiny and spawning a variety of novel professions variously attending to the ethical, fair and compliant information handling and production across all stages of the data lifecycle and application of artificial intelligence.
In 1976, firms comprised of intangible assets held a 16 percent share of the Standard & Poor’s 500 market capitalization; by the 2010s, this had risen to the 80–90 percent range making enforcement of Intellectual Property rights the biggest source of economic rent capture. [4] This latest wave of AI now crashing over present-day societies and cultures, globally is, by its very nature, particularly data-intensive. The larger the model (quantified presently in, “billions of parameters”) and the more data available to train the model (exceeding “trillions of tokens”), the higher the expected performance. This insatiable appetite for data lies in the connectionist, deep learning foundations of the latest AI architectures which apply all this training data to in effect, rehearse (“train on”) the best pattern of responses to vast combinations of input exploring almost every pattern of stimulus. During their training phase, popular Large Language Models are fed varying lengths of text and trained to predict the most likely next word – so called Skip-Grams. Given three words, predict the most likely fourth, given four words, predict the fifth and so on – such that, after training on corpora approaching the entire literary output of humanity across the ages, the model is able to parrot suitably arranged patterns of words in any given language absent any comprehension of what it is saying.
This absence of comprehension surfaces in what today are euphemistically called “hallucinations” – when, tangled up in its complex of parameters, the AI model produces some radically unsuitable or outlandish response. Whether patent absurdity or subtle distortion, such hallucinations represent an unacceptable risk of harm in many applications. A great deal of research is presently aimed at reducing hallucinations and mitigating their risks and harms. Another major area of research seeks to create models that are “explainable”, since, for now the basis of responses and decisions they issue is essentially inscrutable owing to the immense scale and complexity of their architectures and training regimes.
Myriad new career paths will open up for the intrepid lawyer amidst such turbulent industrial and economic upheavals as AI stands to precipitate. Sophisticated intelligent systems and services of all kinds will require talented Business Analysts capable of stewarding the capable, compliant and ethical development and deployment of AI technology. Entire new industry segments are emerging to meet the demand for data protection, digital identity, privacy and related services – with new professional qualifications and designations expanding alongside. For those intrigued by the business and operational opportunities, these intelligent systems pose new risk management challenges demanding the comprehensive and prudent governance well served by a trained lawyer’s mindset at or near the strategic or administrative helm.
Ex judicata staff with support of consulting subject matter experts – Together, we’re following these developments with intense interest in order to support the ex judicata community with valuable and timely access to important learning and development along these exciting and continually multiplying career trajectories. Along with articles to inform and inspire your career evolution, we’ll be sourcing valuable learning and networking resources from and for the network; like online courses and workshops and digital tools and platforms relevant for those moving in these intriguing and rapidly emerging fields.
References
[1] https://inplp.com/latest-news/article/a-brief-history-of-data-protection-how-did-it-all-start/
[3] European Commission Press release, 2022 “Data Act: Commission proposes measures for a fair and innovative data economy “
[4] Dan Ciuriak, 2020, “Economic Rents and the Contours of Conflict in the Data-driven Economy”
Early “data protection” laws and regulations [1]
- 1890: Two United States lawyers, Samuel D. Warren and Louis Brandeis, write The Right to Privacy, an article that argues the “right to be left alone”, using the phrase as a definition of privacy.
- 1948: The Universal Declaration of Human Rights is adopted, including the 12th fundamental right, i.e. the Right to Privacy.
- 1950: The EU Convention on Human Rights sequence of fundamental rights is amended, with articles now appearing in a different order.
- 1967: The Freedom of Information Act (FOIA) comes into effect in the US and gives everyone the right to request access to documents from state agencies. Other countries follow suit.
- 1980: OECD issues guidelines on data protection, reflecting the increasing use of computers to process business transactions.
- 1981: The Council of Europe adopts the Data Protection Convention (Treaty 108), rendering the right to privacy a legal imperative.
- 1983: The Federal Constitutional Court of Germany reaches a fundamental decision regarding the census judgment. The verdict is considered a milestone of data protection.
- 2023: The Data Act – law to enhance the European Union’s data economy by making industrial data more accessible and usable, encouraging data-driven innovation.
Recent Milestones in Artificial Intelligence
- 1997, IBM’s Deep Blue – first computer chess-playing system to beat a reigning world chess champion, Garry Kasparov, under regular chess match conditions
- 2004, NASA Mars rovers Spirit and Opportunity equipped with AI to navigate rocky terrain.
- 2011, IBM Watson DeepQA outcompetes two of US Jeopardy quiz show’s most formidable champions, Ken Jennings and Brad Rutter.
- 2012, Geoffrey Hinton, Alex Krizhevsky and Ilya Sutskever, present groundbreaking visual-recognition network AlexNet to win the ImageNet competition.
- 2011 – 2016, Siri and Alexa respond to a lengthy list of spoken and questions and not answer those outside their purview.
- 2016, Hanson Robotics create Sophia, a “human-like robot” capable of facial expressions, jokes, and conversation – Saudi Arabia granted Sophia citizenship in 2017, making her the first artificially intelligent being to be given that right. (criticized as Sophia thereby granted rights which Saudi women were then still denied)
- 2016, DeepMind’s AlphaGO goes on to beat Lee Sedol, one of the best Go players in the world.
- 2020, GPT-3 – a 175 billion parameter large language model (LLM) able to generate computer code, poetry, and prose – reviewed by Farhad Manjoo in The New York Times, GPT-3 was described as “amazing”, “spooky”, “humbling “, and “more than a little terrifying”.
Life on Another Planet
Life on another planet: Moving from law firm to nonlegal job in Corporate America. <p class="subhead"> <br> Enter a world where questions are actually encouraged, and concepts like ‘teamwork’ are more than a curiosity.</strong>
The Happiness Project*: Lawyer Edition:
Overcoming Common Hurdles to Leaving Law for a New Career. A growing number of lawyers are hitting the wall. That is, coming to the realization that they don’t want to continue to practice law. It’s not just lawyers of course. In the age of the Great Resignation, all sorts of people are realizing they want to seek greener pastures. Even if they don’t know what those greener pastures are.
What Is Your Elevator Pitch?
If you ask 100 attorneys, who successfully transitioned to careers in business, how they did it you’ll likely get 100 different answers.