The EU AI Regulations Are Coming. How Does OpenAI React?
EU AI regulations
Artificial intelligence, or AI, is a powerful technology changing the world. But it also comes with challenges and risks. The European Union, a group of countries in Europe, is working on EU AI regulations.
However, not everyone is happy with these rules, including OpenAI - the leading developer of general-purpose AI systems such as GPT-4. In this article, we will see what is EU AI Act and how OpenAI reacts to these new rules.
Growing Concern About AI
In this digital world, AI development is constantly expanding. People are facing worries and fears about the possible negative effects of AI on our society and lives. That's the reason why AI should be regulated. Besides its excellent benefits, here are some of the main concerns:
- AI bias: People are concerned that AI systems might mistreat certain groups or discriminate against them based on race, gender, age, or other traits. This bias could lead to unequal chances and outcomes for different people.
- AI safety: Making sure AI doesn't harm people, things, or the environment is a big challenge. Many are worried that AI could cause accidents or be used to hurt others on purpose, which could be harmful.
- AI ethics: People are thinking about the right and wrong ways to use AI. They worry about how AI might affect our dignity, privacy, freedom, and rights. They want to make sure AI is used in a good and fair way, especially in areas like healthcare and decision-making.
- AI governance: Having clear rules and standards for AI is crucial. There must be ways to check that AI is being used responsibly and that there are matters if something goes wrong.
These concerns reflect the growing awareness of the potential risks and challenges of advancing AI models. It's vital to think carefully and work together to ensure AI helps us without causing harm.
The EU AI Act
EU AI Act
The AI Act is a law being considered in Europe. It would be the first major law about AI. The new rules about AI follow a system where they consider how risky the AI can be.
They will exclude AI systems that could be dangerous and harm people's safety. This includes systems that try to trick or manipulate people or exploit their weaknesses. It also involves models that classify people based on their behavior or personal details.
Members of the European Parliament (MEPs) made crucial changes to the list of banned AI uses. Some of the things they banned include:
- Using biometric identification systems that can identify people in real-time in public places.
- Using biometric identification systems without permission, except for serious crimes and only with approval from a judge.
- Using systems that classify people based on personal traits like gender, race, religion, or political beliefs.
- Using AI to predict where crimes might happen or who might commit them based on profiles or past behavior.
- Using AI systems to guess people's emotions in law enforcement, borders, workplaces, or schools.
- Collecting biometric data from social media or cameras without permission to create facial recognition databases, which violates privacy rights.
Right now, the European Parliament and the Council of the EU are discussing the proposal. They need to agree on the final rules before they become law.
Just like the GDPR did for data protection in 2018, these rules could become a global standard for AI. They're already getting attention around the world. In September 2021, Brazil's Congress passed a bill to create rules for AI. Yet, it still needs the approval of the country's Senate.
EU AI Regulations' Impact
The EU's new rules for AI will affect companies worldwide. In the past, EU rules on data privacy and phone chargers influenced the rest of the world. However, some think these AI rules could be too strict and limit innovation.
EU AI regulations impact
Besides possible business threats, AI companies have to face other issues. For example, OpenAI uses a lot of data from the internet to train its AI systems. Some of that data is copyrighted.
If they have to reveal where they got the data, OpenAI could be sued by companies who own the copyrights. Another AI company, Stability AI, is already facing illegal issues for using copyrighted data.
How OpenAI Reacts
Some vital people in the tech industry, like Elon Musk and Steve Wozniak, want a pause to think about the risks of AI before making these rules. The CEO of OpenAI, Sam Altman, has said that his company might stop offering its services in Europe because of new EU AI Regulations.
Altman has concerns about these rules and how they will affect their big AI systems like ChatGPT and DALL-E.
EU AI regulations OpenAI react
Altman explained that they would try to follow the rules. Yet, they might have to stop serving in Europe if they couldn't. It's essential to pay attention to the specific details of the rules because they can make a big difference for companies like OpenAI.
Concerns and Timeframes
Even if the AI Act gets approval by the end of the year or early 2024, it won't become effective right away. Companies and corps will have a grace period to figure out how to follow the new rules.
Some people think that the final version of the AI Act goes beyond what was initially proposed. Industry experts might ask for more time to comply with the rules. Instead of having one and a half to two years, they need two to three years to adapt.
EU AI regulations concerns
Frederico Oliveira Da Silva, a legal officer at a European consumer group, pointed out that ChatGPT, for example, was launched only six months ago. However, there have already been both problems and benefits with it.
If the AI Act takes a long time to come into effect fully, people are concerned about what might happen during those years. That's why they want authorities to pay close attention and focus on this technology.
The new EU AI regulations have caused a lot of talk and reactions from many people. Part of them thinks these rules are trustworthy and can protect everyone's rights. However, some people worry that these rules might make it harder for new and innovative ideas to come up. There also might be challenges in following the rules.
As the EU AI regulations are being finalized and put into action, the people in charge must listen to everyone's concerns and thoughts. They must find a good balance that urges AI's responsible use while avoiding unintended problems. These decisions will shape how AI applies in Europe and will also have an impact on AI around the world.