Uncategorized

How and Why Europe Wants to Regulate ChatGPT | technology



All organizations want to organize ChatGPT. A popular chatbot based on generative AI is starting to raise suspicions around the world. There are concerns about its impact on the privacy of users, the possibility of dissemination of misleading information and the possibility of destroying the functions it can cause if it is implemented in the work environment and perform the functions of workers.

Two weeks ago, in a joint letter, more than a thousand professionals demanded that development of these programs be halted for six months. Governments have responded to this call for help. The US opened a two-month period on Tuesday to collect ideas on how to regulate this type of tool. China has already introduced regulations to regulate generative AI, according to Reuters: companies that want to use the technology must demonstrate to Beijing that they meet a series of security requirements.

What is happening in Europe? On the ancient continent, the situation is more complicated. It is progressing at a different pace and in several aspects at the same time. Here are some of the keys to the debate in the European Union these days:

How does ChatGPT work?

ChatGPT4 is the latest version of the Language Large Model (LLM). This is the name given to systems of artificial intelligence, or more specifically deep learning, trained with massive amounts of data (in this case, text) to be able to maintain dialogues with the user. The program processes millions of texts (in ChatGPT4’s case, the entire Internet) and applies a series of algorithms to try to predict which word is most likely to follow previous words in a coherent sentence. For example, if you type “the sky is colored,” that means the system has been trained by reading enough text to be able to say “blue.”

The increase in computational power in recent years, the development of the algorithms responsible for carrying out this training, and the huge databases used in the process have made the process go beyond just predicting a single word. ChatGPT, like other similar templates, is capable of writing coherent full texts without spelling errors. The algorithm takes 175 billion parameters into account each time a question is asked. The result can be amazing.

See also  An out-of-control Tesla kills two people in China | technology

What threat does it pose to users?

The problem with the answers ChatGPT provides is that they are coherent, but not necessarily correct. This is how the program warns itself when it starts: “It can sometimes generate incorrect information,” as well as “harmful instructions or biased content.” It was also warned that the system “has limited knowledge of the world after 2021”, which is the date the database was covered with which the model was trained.

In addition to the quality of the information (or misinformation) that the tool generates, there is a fear that the further sophistication of the model (ChatGPT4 has just been introduced and version 5 is already in the works) will do the job. Until now.

What has been done so far in the European Union?

Spain’s Data Protection Agency (AEPD) and its European counterparts met Thursday at the European Data Protection Committee (EDPB), the body in which they coordinate forces. They decided to set up a “working group” to exchange information on this issue. Community organizations have not adopted any specific measures regarding ChatGPT.

Those who have done so are some Member States. Most recently, Spain: The AEPD announced Thursday afternoon to “commence ex officio preliminary investigation procedures” for OpenAI, the company that developed ChatGPT. According to the agency’s sources, this does not mean that measures will be taken against the company, but rather it is exploring whether the situation “justifies addressing the action.” For its part, Italy blocked the app two weeks ago until it was determined whether it violated European data protection regulations. French, Irish and German authorities are investigating this possibility.

On the other hand, the Internal Market Commissioner, Terry Breton, announced last week that AI-generated content must carry a specific warning of its origin. “In everything that is generated by artificial intelligence, whether it be text or images, there will be an obligation to notify that it has originated from them,” he said.

See also  How to save energy (and money) thanks to smart devices | your technology | Country

Why do you want to organize?

To date, the most immediate threat to the interests of European citizens in terms of privacy has been detected. ChatGPT itself warns its users not to enter “personal data” into the system. But this is not enough.

In response to a question about the reasons that prompted him to initiate “prior investigation procedures”, EPA sources said that they could not reveal the details so as not to harm the handling of the file. According to the experts consulted, potential breaches of the regulations could have to do with the use of conversations that ChatGPT users have with the same tool to train them, as well as the retention of their personal and payment data that does not comply with all the security safeguards required in the EU.

How would you like to organize?

A debate is taking place in Brussels behind closed doors about the best way to handle the situation. The majority view is that the EU already has (or will be given) sufficient regulations to control the potential negative impacts of generative AI.

The spearhead of this strategy is the European Regulation on Artificial Intelligence (Artificial Intelligence Law), a document that has been negotiated since 2019 and whose final text has not yet been approved by all European institutions. It is not expected to enter into force before the end of next year or until 2025. The text classifies technologies according to the risks they pose to citizens and sets restrictions accordingly. Most benign ones can work without problems, but high-risk ones are off limits. They fall into this last chapter, for example, using automatic facial recognition systems in public places or social credit scoring systems, which is something that already works in China.

See also  CNMV does not see enough signs of coordinated action to control Indra, but will take measures so that this does not happen again

During the French rotating presidency of the European Union, a recommendation was made to include generative AI in the high risk category. “It seems reasonable to me, but there is a lot of pressure from tech companies not to be included there,” an MEP who was involved in negotiating and drafting the AI ​​regulation told EL PAÍS.

Another MEP involved in the process, Romanian Dragos Todorac, said at a conference this week that the EU’s response to ChatGPT challenges should be guided by the regulation of AI law, not by the General Data Protection Regulation. Regulations that Italy adhered to to ban the tool. “I am convinced that we need a unified response,” he stressed.

Is a new regulation needed?

The committee’s sources consider future AI regulation “designed to test future challenges.” By regulating the uses, not the technologies themselves, regulating can adapt to challenges like the sudden explosion of the generative AI phenomenon, which we’ve been all too familiar with since ChatGPT’s open launch in November of last year.

Under this approach, the technology behind ChatGPT will not be banned unless it is considered high risk. “If someone uses it, for example, to process health data, then the regulations will apply,” says Jan Philipp Albrecht, head of the German Heinrich Böll Foundation, which is linked to Los Verdes. “This is the problem that is emerging now: if we want to regulate generative AI more strongly, a new category must be introduced stating that general-purpose AI should be considered high risk if it can mediate risk domains,” he adds. This is what tech companies want to avoid at all costs; For this they organize meetings with European legislators.

You can follow country technology in Facebook And Twitter Or sign up here to receive The weekly newsletter.

Subscribe to continue reading

Read without limits



Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button