Two reports out this week look at how large language models (LLMs) – deep learning algorithms that can recognize, summarize, translate, predict, and generate content using very large datasets – could impact the future of work.
A World Economic Forum white paper says that LLMs could be a boon for jobs that require critical thinking, complex problem-solving skills, and creativity, including those in engineering, mathematics, and scientific analysis. These tools could benefit workers by increasing the productivity of routine tasks and making their roles more rewarding and focused on a higher added value but a Boston Consulting Group (BCG) report cautions that GenAI is a double-edged sword.
When GenAI has mastered a type of task, virtually everyone (about 90%) will see a boost in performance, says the BCG report. The trouble is people find it difficult to determine when GenAI has achieved this mastery. This is critical, according to the report, because GenAI can create significant value (40% performance improvement over not using GPT-4) for creative product innovation, but also erode value by as much as 23% when used for business problem solving.
The BCG report found that in some cases GenAI can be good enough to deliver final drafts but even if used for the right task GenAI may hurt a company’s overall creativity.
While a lot of the discussion around GenAI has focused on using the technology for first drafts and then augmenting it with human intelligence, the BCG report found that if people make changes to GPT-4 output on tasks the technology has mastered they tended to distract from its quality: For every 10% increase in divergence from GPT-4’s draft, participants dropped in the quality ranking by approximately 17 percentile points. People should instead focus their efforts on tasks that are still beyond GenAI’s mastery, according to the BCG report.
The report also adds another note of caution: Use of GPT-4 tends to make output considerably more homogenous, reducing diversity of ideas at the group level by as much as 40%, potentially impacting a company’s creativity.
While the BCG report focused on how people can create – and destroy- value with GenAI, the Forum’s white paper focuses on LLM’s potential impact on specific sectors and jobs.
According to the Forum’s analysis, which examined more than 19,000 distinct tasks across 867 different occupations likely to be impacted by LLMs, the industries with the highest estimates of overall potential exposure – both in automation and augmentation – are financial services and capital markets, along with insurance and pension management.
As LLMs advance, new roles will also be created, including AI developers, interface and interaction designers, AI content creators, data curators and specialists in AI ethics and governance.
The jobs most at risk of automation – with up to four-fifths of the tasks automated – are those that involve routine and repetitive language tasks, including roles such as credit authorizers, checkers and clerks. The occupations projected to remain relatively unaltered include education, guidance, career counselors and advisers, with 84% of their tasks having a low exposure to change, according to the white paper, which reinforces the findings from the Forum’s Future of Jobs Report 2023, which found that up to a quarter of jobs are expected to change in the next five years from the combined impact of technology, the green transition and the global economic outlook.
“Generative AI is poised to impact labor markets significantly, but this impact will be highly different across different roles,” Saadia Zahidi, Managing Director, World Economic Forum, said in a statement. “Business leaders, policy-makers and employees must collaborate on harnessing the potential of new jobs while managing displacement and ensuring a future of work that empowers and elevates people.”
IN OTHER NEWS THIS WEEK
SUSTAINABILITY
Climate Takes Center Stage As Global Business And Political Leaders Meet In NYC
With the world on track to break the record for the hottest year in history, world leaders, business leaders, celebrities and activists converged on Manhattan for Climate Week and the U.N.’s Climate Action Summit, to focus the world’s attention on the climate crisis. The annual climate gathering, along with the World Forum Sustainable Development Impact Meetings 2023, coincided with the start of the United Nations General Assembly, bringing heads of state and top government officials together with private-sector leaders to focus on climate change in a year marked by a record number of billion-dollar disasters, including eight severe floods
A report published this week by Climate Impact Partners on the climate commitments of the world’s largest companies explores why climate pledges alone are not delivering enough real, immediate change. Despite a steady rise in climate commitments following the Paris Agreement in 2015, in the last year the number of Fortune Global 500 companies with a significant climate commitment has stagnated at around 66% of companies, says the report. While commitments are stagnant, emissions are still on the rise. But reducing emissions pays off both environmentally and financially: companies that reduced reported emissions year over year earned on average nearly $1 billion more in profit per company than their Fortune Global 500 peers, according to the report.
The Forum and its partners launched or advanced more than 25 initiatives during the Sustainable Impact meeting in NYC focused on driving immediate and long-term progress, including progress on key initiatives such the First Movers Coalition, the EDISON Alliance and the Reskilling Revolution, ahead of the Forum’s Annual Meeting 2024 in Davos.
The initiatives include:
- A fund to strengthen food value chains in Africa through innovative financing and better support for small and medium agricultural enterprises announced by the U.S. Agency For International Development (USAID) and the Norwegian governme,
- A Humanitarian Investing Initiative call to action to mobilize $10 billion for businesses in frontier markets to support vulnerable communities signed by more than 50 organizations.
- A new Forum report showed how to apply technology and reduce regulatory challenges to build a cross-border payment system that helps realize the benefits of international trade.
- 61 global experts from the Forum’s Network of Global Future Councils shared their insights, covering 23 distinct thematic areas with an interdisciplinary approach to generating fresh ideas and innovative thinking on how to achieve the SDGs.
- In an open letter, corporate members of the Global Alliance for Social Entrepreneurship, hosted by the Schwab Foundation, expressed their strong support for the implementation of the UN resolution on promoting the social and solidarity economy.
- Advanced manufacturing CEOs proposed seven principles for facilitating a coherent and seamless transition in the wake of disruptions in a new Forum white paper.
- The Forum introduced a new Polar Tipping Points hub in its Global Collaboration Village. The platform, which is powered by next-generation technology, will bring global leaders together in the metaverse to solve the deepening effects of the climate and nature crisis. The hub is an immersive virtual reality space and will help confront the pressing issue of climate tipping points, particularly those located in the polar regions.The hub offers an immersive experience that allows collaborators to visualize the ramifications of polar warming. Through a monitoring station, visitors gain a deeper insight into three of the five pivotal polar tipping points that become precarious if temperatures rise by 1.5°C or beyond.
QUANTUM COMPUTING
Quantum Machine Learning Solution Offers Faster Routing In Disaster Situations
Over the past five decades, disaster occurrences have amplified fivefold, propelled by climate change, extreme weather patterns, among other factors. In disaster scenarios, a timely response is vital and significantly impacts public safety. However, the dynamic and unpredictable nature of these situations poses a severe challenge for technical solutions.
Hybrid quantum technologies look poised to come to the rescue.
Terra Quantum, which offers corporates quantum algorithms, software that can run today on classical computers, while the development of quantum hardware proceeds, and Honda Research Institute Europe (HRI-EU), a center for advanced research in intelligent systems, are collaborating to develop new methods to shorten escape times in emergency situations.
As part of the Proof of Concept (POC), an earthquake scenario was simulated on a realistic map of a small town. A hybrid quantum solution demonstrated promising results, predicting efficient dynamic escape routes for vehicles, and minimizing evacuation times.
“Identifying realistic problems where quantum technologies may unfold their potential constitutes one of the biggest challenges in the field today,” Dr. Sebastian Schmitt, Principal Scientist at HRI-EU said in a statement. “This work represents a promising step in that direction and shows how to employ hybrid quantum-classical learning architectures in a real-world use-case.”
The developed quantum computing solution can account for numerous real-time variables and shows competitive efficiency results compared to traditional computing methods. It utilizes quantum machine learning and requires only local information to make decisions, which amounts to less than 1% of the map information — a crucial factor in an evolving and uncertain emergency scenario. The solutions are obtained from quantum simulations performed on classical computing hardware. This application can be extended to be executed on large-scale quantum computing hardware in the future.
This is just the start, we see the huge potential in applying quantum technologies in the automotive and mobility sectors,” said Markus Pflitsch, founder and CEO at Terra Quantum. (For more on Terra Quantum see The Innovator’s recent Startup Of The Week article)
HEALTH
DeepMind, Meta Announce New Separate Efforts To Apply AI To Human Health
Researchers at Google DeepMind have used artificial intelligence to predict whether mutations in human genes are likely to be harmful, in one of the first examples of the technology helping to accelerate the diagnosis of diseases caused by genetic variants. The AI tool, called AlphaMissense, assessed all 71 million ‘‘missense” mutations, in which a single letter of the human genetic code changes. Of these, 32% were classified as likely to be pathogenic, 57 % benign and the rest uncertain. The findings were published on Sept, 19 in the journal Science.
Meanwhile, Meta chief executive Mark Zuckerberg announced Sept. 19 that the philanthropic group he founded with his wife Priscilla Chan, the Chan Zuckerberg Initiative, would build “one of the largest computing systems dedicated to non-profit life sciences”. It will focus on using AI to study human cells and diseases.
Musk’s Neurolink To Start Human Trial Of Brain Implant For Paralysis Patients
Billionaire entrepreneur Elon Musk’s brain-chip startup Neuralink said September 19 that it has received approval from an independent review board to begin recruitment for the first human trial of its brain implant for paralysis patients. Those with paralysis due to cervical spinal cord injury or amyotrophic lateral sclerosis may qualify for the study, it said, but did not reveal how many participants would be enrolled in the trial, which will take about six years to complete
FINANCIAL SERVICES
Rival Banks Unite To Take On Big Tech
America’s largest banks have announced their latest effort to defend their turf from Big Tech groups. JPMorgan Chase, Bank of America, Wells Fargo, and others, next year plan to launch Paze: a mobile wallet that will connect directly to the credit and debit card accounts of 150mn customers. The app will be operated by Early Warning Services, a bank consortium group that already runs payments app Zelle. Paze is the latest sign that big banks see partnerships — either working collectively or even in collaboration with tech firms — as the best way to stop the advances of the likes of Apple, Google and, most recently, Elon Musk’s X (formerly known as Twitter), which aim to offer banking services to their millions of users.
ARTIFICIAL INTELLIGENCE
Chat GPT Maker Open AI Accused Of Violating Europe’s Data Privacy LAws
Techcrunch Europe reports that OpenAI is facing another investigation into whether its generative AI chatbot, ChatGPT, complies with European Union privacy laws. Last month a a complaint was filed against ChatGPT and OpenAI in Poland, accusing the company of a string of breaches of the EU’s General Data Protection Regulation (GDPR). On September 20 the Polish authority took the unusual step of making a public announcement to confirm it has opened an investigation.
Top U.S. Authors Sue OpenAI Over Copyrights
Reuters reports a trade group for U.S. authors has sued OpenAI in Manhattan federal court on behalf of prominent writers including John Grisham, Jonathan Franzen, George Saunders, Jodi Picoult and “Game of Thrones” novelist George R.R. Martin, accusing the company of unlawfully training its popular artificial-intelligence based chatbot ChatGPT on their work. The proposed class-action lawsuit filed late on Sept. 19 by the Authors Guild joins several others from writers, source-code owners and visual artists against generative AI providers. In addition to Microsoft-backed OpenAI, similar lawsuits are pending against Meta Platforms and Stability AI over the data used to train their AI systems.
To access more of The Innovator’s News In Context articles click here.