AI Index Report 2024: Transforming industries, workforce; exposing job displacement woes

The integration of AI across various industries has transformed the way we live and work. Fueled by rapid technological advancements, AI is said to have boosted efficiency, accuracy, and productivity, even accelerating scientific discovery.

Released on April 15, the 2024 AI Index Report from Stanford University’s Human-Centered Artificial Intelligence Institute (HAI) highlights the progress, with AI models now capable of exceeding human performance in several benchmarks, including image classification, visual reasoning, and English comprehension.

The report noted a critical development in AI capabilities over the past year. The “state-of-the-art” models like Gemini, GPT-4, and Claude-3 demonstrated impressive abilities in generating “fluent text, processing audio, and even explaining memes.”

Google’s Gemini Ultra became the first Large Language Model (LLM) to achieve human-level performance. The analysis was conducted through the Massive Multitask Language Understanding (MMLU) benchmark, which tests performance across 57 scenarios in humanities, science, and social sciences. 

Meanwhile, OpenAI’s GPT-4 achieved an average success rate of 0.96 on the Holistic Evaluation of Language Models (HELM) benchmark, which tracks average performance across various tasks.

Investments in AI

AI’s swift progress stemmed from ongoing technological advancements coupled with growing human engagement. 

Additionally, it is fueled by the soaring global private investment of $25.2 billion in 2023 from $3 billion in 2022 in major generative AI companies like OpenAI, Anthropic, Hugging Face, and Inflection, as stated in the report.

The United States led the world in total private investment in AI in 2023, with $67.2 billion surpassing China’s $7.8 billion by a factor of roughly 8.7. This significant funding gap underscores the widening lead that the U.S. maintains in AI research and development.

The recent boom, as mentioned in the report, drove up the cost of powerful computer chips needed for training and running complex AI programs.

OpenAI’s GPT-4, for example, required an estimated $78 million in computing power, while Gemini Ultra cost a staggering $191 million.

The Washington Post also noted that big tech companies could become even more powerful due to a surge in investments in AI. Investors are channeling hundreds of millions of dollars to help these firms develop proprietary AI tools, the report says.

This influx of funding has facilitated the training of increasingly complex AI models to deliver productivity gains and efficiency improvements across various industries.

AI, the economy, and the workforce

Advancements in AI significantly contributed to worker productivity and efficiency, leading to higher-quality outputs. Data from the report revealed increasing AI integration across the economy has enabled workers to complete tasks faster and produce better work.

According to an August 2023 study from the ICPA Journal of Business Administration and Economics, integrating AI into various industries could bring increased accuracy, reduced costs, and improved customer satisfaction for businesses and consumers. This is true given AI’s ability to automate routine tasks, improve decision-making processes, and boost labor productivity and output accuracy.

The 2024 AI Index Report predicted a potential revenue increase of 4.8% to 9.3% for the high-tech industry due to generative AI adoption. 

Beyond the revenue gains, businesses are reaping the benefits of this technological advancement with revenue increases and cost reductions.

In a recent McKinsey survey, costs have dropped by 10%, indicating AI is generating “significant efficiency gains for businesses.”

However, the expanding role of AI in the global economy has also raised concerns about job displacement and the changing nature of work.

This concern was echoed by a TechSpot report on the case of Suumit Shah, CEO of the Bengaluru-based company Dukaan. Shah replaced 27 members of his customer support staff with Lina, an AI-powered chatbot trained to handle most customer inquiries, which is a lot cheaper than paying the workers. 

Many companies employ millions of people in call center positions, especially in India and the Philippines. There is a greater impact and fears among these workers when cheaper generative AIs could replace them sooner than anticipated.

A recent Ipsos survey found that 57% of respondents believe AI will alter how they perform their current jobs within five years, while 36% fear AI could replace them entirely in that timeframe. 

Concerns about job displacement due to AI are rising alongside the potential for declining employment, particularly in knowledge-intensive industries. 

The McKinsey survey emphasized that these industries are likely to experience greater disruption but potentially reap more value from AI adoption.

The AI Index Report 2024 showed a decline in employment for specific roles, such as service operations (54%), supply chain management (45%), and Human Resources (41%).

The displacement of workers, especially those with limited job skills and education, can worsen existing inequalities in the job market. This could lead to a decline in labor demand, lower wages, and fewer hiring opportunities. 

Therefore, these challenges underscore the need for effective regulation and transparency from AI companies to mitigate the risks and limitations of advanced AI models.

Responsible AI and regulation

As the advancements in AI continue to transform our world, concerns over the potential risks of AI are also escalating.

Algorithmic discrimination, data privacy breaches, and existential threats have emerged as pressing issues demanding comprehensive regulatory measures.

The AI Index has shed light on the growing efforts by governments and policymakers to address these challenges. The report noted an increase in the number of US agencies issuing AI-related regulations, rising from just one in 2016 to a staggering 25 in 2023.

The report also examined 32 countries that have enacted at least one AI-related bill. The Philippines, which has passed five AI-related bills into law, and Slovenia, which has passed three, were both mentioned in the AI Index Report.

This surge in regulatory activity underscores the urgency to establish clear guidelines and safeguards for the responsible development and deployment of AI technologies.

Apart from the need for more regulation, the AI Index stressed the challenges in comparing the risks and limitations of leading AI models. 

While developers such as OpenAI, Google, and Anthropic utilize responsible AI benchmarks, a lack of standardization makes comparisons difficult, according to the evaluation.

The evolving landscape of AI poses a challenge in crafting effective regulations due to the lack of standardization.

This challenge is further amplified by a concerning rise in AI misuse, with the AI Incident Database recording a 32.3% increase in incidents since 2022. 

For instance, there has been a rise in AI-generated deepfakes of celebrities, safety concerns regarding self-driving cars, and privacy issues related to romantic AI chatbots.

Furthermore, the report underscored that some organizations in Europe, North America, and Asia have implemented measures in areas such as fairness, transparency, privacy, security, and governance to mitigate and address the potential risks of AI.  – with reports from Angelica Paller/Rappler.com