While artificial intelligence (AI) has emerged as the next major wave of innovation, its power dynamics are not evenly distributed.
This is according to Mozilla’s Internet Health Report 2022, which examines how humanity and the internet intersect, scrutinising the nature of an AI-driven world.
AI in this case includes a wide range of automation and algorithmic processes, including machine learning, computer vision, natural language processing, and more, states the report.
Mozilla found that the growing power disparity between who benefits from AI and who is harmed by AI is the top challenge facing the health of the internet.
Solana Larsen, Mozilla’s internet health report editor, explains: “The centralisation of influence and control over AI doesn’t work to the advantage of the majority of people. We need to strengthen technology ecosystems beyond the realm of big tech and venture capital start-ups if we want to unlock the full potential of trustworthy AI.”
Research and advisory firm IDC forecasts that worldwide revenue for the AI market will grow by 19.6% year-over-year in 2022, to $432.8 billion, with the market expected to break the $500 billion mark in 2023.
Mozilla’s report highlights that from social media feeds to fast food restaurants, companies in every sector are turning to AI to unlock new ways to collect and analyse data to tailor their offerings.
However, the benefits and harm are not evenly distributed, it states.
“The companies with resources to invest are carving out competitive advantages. And the countries with access to engineers, large amounts of data and computational power are consolidating their dominance of software and hardware in ways that impact how AI is deployed worldwide.
“The US and China are far ahead when it comes to private investments in AI. That is just one indicator of how differently the rise of AI is experienced worldwide.”
Private investments in AI by country in 2021
Furthermore, the AI power imbalance is reflected in research and development, according to the report.
It reveals the cost of training machine learning systems has decreased, and the availability of data is greater. As more of the world delves into AI, a major imbalance is reflected in the landscape of AI research papers, it states.
“In thousands of papers, it is the same datasets from just a few countries that are used most often to evaluate the performance of machine learning models everywhere.”
For example, the report shows that datasets originating in the US account for the most usages (26 910). Meanwhile, there’s no usage of datasets from African nations, except Egypt.
According to the report, this doesn’t mean that datasets or machine learning models aren’t being developed in the rest of the world.
“The discourse about how AI should be used – and who should benefit from it – is currently heavily weighted toward people and institutions who already wield tremendous power over the internet and the world.
“In fact, more than half of the datasets used for AI performance benchmarking across more than 26 000 research papers were from just 12 elite institutions and tech companies in the US, Germany and Hong Kong (China).”
It further states that a large and frequently reused dataset does not guarantee better machine learning than a smaller one designed for a specific purpose.
“Machine learning models and datasets reflect both the biases of their makers and power dynamics that are deeply rooted in societies and online, but this is not widely acknowledged. More datasets should be created specifically to diversify machine learning methods for equity.”
To build AI trust, the report recommends new techniques and methods for preserving privacy, for logging the origins of data, for operationalising ethics, for auditing algorithms and giving users more power.
“Regulation can help set guardrails for innovation that diminish harm and enforce data privacy, user rights and accountability. Many laws already apply to AI, but policies that are specific to AI (or sometimes bans) are also surfacing in different regions, countries and cities.
“Collaboration across multiple sectors is necessary for solutions,” it concludes.
Share