It always seemed like science fiction: cars that drive themselves without us humans having to do a thing. We all know by now that this is no longer science fiction, but that it’s actually within reach. Now it’s becoming a reality in our country too: the Netherlands will begin testing self-driving cars starting in 2027. But how does that actually work? That’s what we’ll be discussing in this article!
Why is the Netherlands suddenly testing self-driving cars?
It might seem to come out of nowhere, since our country hasn’t exactly been focusing much on this lately. The government would like to start testing autonomous vehicles in a few years, because otherwise they fear that manufacturers will conduct more testing abroad. This could result in self-driving cars being less attuned to Dutch roads. After all, roads in the Netherlands are generally much busier than roads in other countries. If we want to drive autonomous cars here, it’s important that those cars take that traffic into account.
Changes to the law in 2022
Since 2022, the Netherlands has been required to allow testing of self-driving cars. This is due to European regulations adopted that year. These regulations enable automakers to apply for approval of autonomous vehicles. To do so, the vehicles must first be tested on public roads; otherwise, they cannot be approved. In addition, government agencies and car manufacturers want to be able to test self-driving cars on public roads for other reasons. In this case, for example, it concerns models that are not yet ready for type approval. By testing these vehicles on public roads anyway, you also immediately test the interaction between humans, the vehicle system, and the road!
What makes Dutch roads so different?
Why is it actually so important for the Netherlands that self-driving cars also undergo testing on our roads? Both car manufacturers and government agencies want to know how the autonomous vehicle system interacts with typical Dutch traffic conditions. Things like road markings and driver behavior tend to vary quite a bit from country to country. That makes it necessary to test the system in each country to ensure it can be deployed safely. In Germany, for example, this would be important because drivers there are generally allowed to drive faster than in neighboring countries. Another benefit of these tests is that road authorities gain insight into what a road must meet to be suitable for self-driving systems.
The Netherlands is testing self-driving cars, but is that really safe?
We can imagine that testing autonomous cars on public roads sounds very unsafe to some people. Car manufacturers and government agencies naturally take the necessary precautions to ensure testing remains safe for other road users. They do this with the help of a so-called “safety driver.” This means that an additional driver is either physically present or remotely monitoring the actions of the self-driving system. Additionally, testing on public roads may only begin after the system has successfully completed all other forms of testing without any issues or areas for improvement. So you don’t need to worry that the tests will increase the risk of accidents!
The hype surrounding artificial intelligence
Many tech companies currently claim that artificial intelligence will eventually surpass the human brain and be capable of doing everything we cannot. Ask Google, OpenAI, or any other tech company deeply involved in AI, and they will tell you that it is inevitable that this technology will one day be smarter than we are. Many people have even expressed concerns about the development of this technology and what it could be used for, including Elon Musk. As a result, most people are convinced that technology will one day become smarter than humans.
AI will never be smarter than humans
According to scientists at Radboud University, that is certainly not the case. They suspect that the current hype surrounding AI is leading to misunderstandings about what is and isn’t possible with this new technology. Their new publication, titled “Reclaiming AI as a Theoretical Tool for Cognitive Science,” explains why claims about artificial intelligence are exaggerated. Their findings show that there will never be enough computing power to create “artificial general intelligence” that delivers the same cognitive performance as the human brain. In addition, the article explains why pursuing this goal is a futile endeavor and a waste of energy resources and raw materials. After all, some of these are already scarce at the moment!
Collaboration among various universities
The publication is a collaboration between several universities, including Radboud University, Aarhus University, the University of Amsterdam, the University of Bristol, and Memorial University of Newfoundland. All researchers who contributed to the publication are experts in the fields of cognitive science, neuroscience, computer science, and philosophy. People often fail to realize that cognitive science is crucial for understanding claims about the capabilities of AI. People tend to overestimate what computers can do and underestimate what our own brains are capable of. According to the researchers, it is important for the general public to gain more knowledge about AI. Currently, everyone believes the claims made by major tech companies, but that is not a good thing. The goal of the universities is to build a better understanding of AI systems so that people can continue to view the tech industry with a critical eye.
The use of AI in the cultural sector
Artificial intelligence opens up a world of possibilities, including within the cultural sector. The heritage sector is already making extensive use of AI, for example to make handwritten letters searchable, for facial recognition in audiovisual material to make it easier to find specific footage, and to establish new connections between European archives. But AI can, of course, also be used for many purposes in the cultural sector:
- Adding value for audiences or organizations: AI can help make cultural offerings more accessible to the public. This can be achieved, for example, through subtitles or automatic translations. Organizations can use AI to accurately estimate visitor numbers and adjust their staffing levels accordingly, which can lead to significant cost savings.
- Automating and streamlining work processes: With artificial intelligence, an organization can automate many work processes, leading to cost savings and greater efficiency. Examples include automatically translating flyers, editing videos and photos, and answering customer questions via a chatbot. Many cultural institutions are understaffed and need to watch their costs, so this would be a great solution!
- Introducing new products: Artificial intelligence is a valuable tool for creating new products. Creators can use it, for example, as inspiration when composing new music, digitally restoring a painting, or identifying new connections within their own collection to generate fresh content.
- Enhancing the audience experience: The cultural sector can use AI to better serve its audiences, leading to a more engaging experience. Examples include personalized digital interactive tours and guided tours using apps, as well as personalized offers on cultural institutions’ platforms.
What negative impact does AI have on cultural development?
We’ve now highlighted many of the benefits that artificial intelligence brings to the cultural sector, but there are, of course, a number of challenges as well. For instance, the quality of AI output is sometimes still questionable, especially since we often don’t know exactly where the information comes from and whether it’s accurate. In addition, creators need to be mindful of copyright, even if you’re using an AI sample solely as inspiration. After all, AI can extract information from a work that is protected by copyright. And the biggest challenge we currently face with the use of artificial intelligence is that it often operates based on biases.
AI and Bias
Ultimately, its use can therefore reinforce certain biases—an effect we obviously do not want from AI, as it often hinders cultural development. For example, many generative AI tools have an American and European bias. This means that the content generated by the AI is often not representative of people from around the world. The way AI is trained determines the results it produces. If only information containing biases about a certain race or gender is used, there is a high chance that the AI will produce discriminatory results. This is one of the reasons companies are still hesitant to use AI in their recruitment process. Fortunately, steps are already being taken to address this challenge, such as training AI with more diverse datasets. So in the future, we can make artificial intelligence bias-free too!