Guy Ben-Ishai is a leading expert in the field of technology economics and innovation.
He is currently the Head of Economic Policy Research at Google, where he oversees initiatives to study and shape public policies for emerging technologies.
Prior to that, Guy Ben-Ishai was chief economist at the New York State Attorney’s Office and principal at Brattle Group, where he focused on competition policy, financial markets, consumer protection and related topics.
He holds a bachelor’s degree in economics from UCLA and a doctorate from the Hebrew University of Jerusalem.
He has also worked as a postdoctoral researcher at the University of Toulouse and was a guest researcher at the Milken Institute.
The discussions focused on the impacts of adopting artificial intelligence on the job market, worker skills, and training. The exchanges revealed both the opportunities AI offers and the challenges to anticipate in order to support a smooth adoption in a context of increased global competition.
One of the questions relating to the use of AI at work concerns the impact of task automation on employment. While AI could automate between 25% and 30% of tasks in a wide variety of professions, Dr. Ben-Ishai believes that this would unlikely lead to mass unemployment. Indeed, the replacement of tasks does not necessarily lead to the systematic elimination of jobs. New tools tend to support human work rather than replace it entirely.
Rather than replacing jobs, the adoption of AI in companies can enhance human skills that cannot be easily automated. By reserving automation for certain tasks for which the machine is more efficient than humans, such as pattern recognition, massive data processing, and prediction, humans can focus on activities requiring their own skills, such as critical thinking and decision-making.
This redistribution of tasks could bring real added value to human-machine collaboration, through a division of labor between the human worker and the artificial agent. This new distribution of human-machine tasks also brings challenges, as it will be necessary to determine where and how to evolve tasks, work processes, and employee skills.
In summary, Dr. Ben-Ishai highlighted three main categories of AI technology applications:
AI, and in particular certain generative AI technologies, could disrupt learning methods by allowing students and workers to access information and skills previously inaccessible without several years of sometimes costly studies.
During this hearing, Dr. Ben-Ishai pointed out that in the United States, a major issue lies precisely in the importance given to university degrees, which condition integration into the labor market and ultimately lead to socio-economic difficulties for those who do not have them.
It is ultimately possible that AI will lower the barriers to entry for certain positions for employees who have not been able to benefit from higher education, by giving them access to learning tools and decision support. One of the side effects to watch out for could be a reduction in the interest of certain higher education degrees in this context of more horizontal access to knowledge.
The adoption of AI systems generates multidimensional effects that manifest themselves both economically and societally. We explored some of these transformations by addressing, on the one hand, topics relating to the impact of AI on the global economy, particularly in terms of productivity and competitiveness, and on the other hand, its capacity to help address some of the key societal challenges of the 21st century, reform the future of work, and advance scientific discovery in a multitude of disciplines. These two components, macro and meso, illustrate the complexity of the changes brought about by AI, highlighting the need for thoughtful and balanced adaptation to these new uses.
Concerns are emerging about the disruption of the world order and whether AI will be inclusive or dominated by a few major players, particularly Western ones.
Dr. Ben-Ishai pointed out that a third of humanity has been connected to the internet over the past fourteen years, mainly in emerging countries, highlighting the rapid growth of users of digital services in this region. He draws attention to the challenges faced by people in these countries in taking full advantage of digital connectivity, stressing the need for access to affordable devices and relevant content.
Furthermore, according to Guy Ben-Ishai, any country wishing to lead the AI race or even simply ensure that AI is implemented on its territory must develop its digital infrastructure. The latter is no longer limited to simple bandwidth. It is now a question of accessing data centers and investing in computing power.
Finally, concerns about the increased computing power required for AI were raised by members of our working group. This could lead to increased power consumption despite efforts to manage this consumption and reduce emissions.
Several participants expressed their concerns about the fears surrounding AI, particularly concerning issues related to bias, disinformation, respect for privacy, or freedom. A recurring concern is the influence of these technologies on education and creativity.
The discussions highlighted the need for thoughtful conversation on this topic. They also highlighted the need to explain to the public that changes due to technology do not happen quickly and are the result of a long process. These discussions can both build trust and collectively apprehend these risks.
The discussion highlighted the complexity of accountability in technology decision-making, stressing the importance of clarifying who is responsible for the consequences of technological advances. Guy Ben-Ishai's remarks also underline the importance of identifying the AI stages of production where such concerns may be addressed rather than accelerating without safeguards towards widespread adoption. These issues are shared by workers and unions in Europe, who express significant concerns about the impact of generative AI, including fears related to job substitution and working conditions.
The discussion also highlighted the importance of employer responsibilities in the context of AI application, emphasizing the fact that a company's reputation is increasingly linked to its ethical actions and how it interacts with its stakeholders.
Finally, Guy Ben-Ishai mentioned the risks of rapid technological acceleration exceeding the adaptive capacity of human beings, insisting on the need for companies to invest in both transformative applications and the ability of their workforce to adapt. In addition, challenges related to the integration of AI in certain key areas for society remain and will have to be addressed on a case-by-case basis. This is particularly the case for health systems, especially with regard to ensuring effective communication and coordination between the various functions within a hospital or healthcare provider, and issues relating to accountability in the medical context.
Nevertheless, he also expresses concern about public mistrust of AI and its potential impact on adoption, particularly in high-stakes professions such as healthcare. "Extreme fears" could be a barrier to innovation.
One of the recurring questions when new technologies emerge is how best to frame their adoption. Our exchanges highlighted certain essential issues to be taken into consideration when setting up a regulatory framework linked to the adoption of AI technologies in professional practices.
The discussion emphasized the need to adopt a balanced approach to AI regulation, highlighting the importance of protecting society while taking advantage of the economic potential of these technologies. Regulation that does not oppose innovation, but rather knows how to establish a protective framework without inhibiting creativity and the adoption of new technologies. There is indeed a fear that excessive regulations will stifle innovation in AI.
Dr. Ben-Ishai also raised the issue of future regulation requiring employers to clarify their positions on the use of technology. This is linked to concerns about replacing employees due to the productivity gains brought about by artificial intelligence. But the facts do not corroborate these concerns: in 2016, a forecast announced that radiologists would be replaced thanks to advances in AI. However, the number of radiologists in the United States has rather increased, thus exposing the complexity of issues related to job displacement.
A major concern has been expressed regarding the need to implement a data management strategy. Guy Ben-Ishai pointed out that while there is a strategy for infrastructure and computing power, a coherent approach to data management is lacking.
It was also noted that the majority of language models are trained on publicly available data collected on the Internet, without being limited to a specific geography or country, which implies the need for a global approach.
.png)
.png)
.png)