Powerful artificial intelligence systems can bring enormous benefits to society and help tackle some of the world’s biggest problems. Machine learning models are already playing an important role in diagnosing diseases, accelerating scientific research, increasing economic productivity, and reducing energy use by optimizing the flow of power in the grid.
It would be a tragedy if those benefits were jeopardized as a result of the backlash against technology. But as one of his “Godfathers of AI”, Geoffrey Hinton, warned last month when he resigned from Google, the abuses of AI technology in areas such as unfair discrimination, disinformation, and fraud. As it increases, the danger increases. It is therefore imperative that governments act quickly to regulate technology appropriately and proportionately.
How we do this will be one of the greatest governance challenges of our time. A machine learning system that can be deployed to millions of use cases can ignore simple classifications and cause many problems for regulators. This rapidly evolving technology can also be used in large-scale, diffuse, invisible and ubiquitous ways. Encouragingly, however, regulators around the world are finally starting to address the issue.
Last week, the White House summoned the bosses of the biggest AI companies to survey the technology’s benefits and dangers before outlining future guidelines. The EU and China are already well advanced in creating rules and regulations to govern AI. The UK competition authority is also due to conduct a review of the AI market.
The first step is for the tech industry itself to agree on and enforce some common principles of transparency, accountability and fairness. For example, companies should not try to disguise chatbots as humans. The second step is for all regulators in areas such as employment law, financial and consumer markets, competition policy, data protection, privacy and human rights to amend their existing regulations to take into account the specific risks posed by AI. is to Third, government agencies and universities develop their own technical expertise to reduce the risk of industrial takeovers.
Beyond that, there are two overarching regulatory regimes to consider when it comes to AI. Even if neither alone is sufficient for the scale of the challenge. One regime based on the precautionary principle means that algorithms used in several critical life-or-death sectors, such as healthcare, the justice system, and the military, require pre-approval before use. This could work much like the US Food and Drug Administration. The U.S. Food and Drug Administration has broader powers to screen drugs before they are released to the public and to protect and promote public health.
A second model, which is more flexible, is based on ‘contingent governance’, as is done in the airline industry. As amazing as this is, it has worked very effectively to raise aviation safety standards over the last few decades. International aviation authorities have the power to order all aircraft manufacturers and airlines to make changes when a fault is detected. Similar methods may be used when harmful flaws are found in consumer-facing AI models, such as self-driving cars.
Several industry-leading researchers are calling for a pause in the development of cutting-edge generative AI models. But a moratorium is pointless unless a clearer governance structure is put in place. Even the tech industry has acknowledged the need for clearer rules and should work constructively with governments and civil rights groups to help create them. After all, a car can go faster through corners if it is equipped with effective brakes.
