Open the gates of the gate bay, Congress

Artificial intelligence is no longer just a tool; It has become collaborative. With the emergence of thinking models and artificial intelligence agents who work independently, we are now entering the era of research and automatic development, as artificial intelligence models are not only responded to information from users, but increasingly, and even in some cases, replace the tasks of technician employees in the leading AI companies.

This transformation may be the most important technological development since the Internet. But unlike previous technology waves, the systems we are now creating learning faster than we can organize, adapt, or even completely understand.

Ask AI’s higher question today, and will receive an answer that has been manufactured from trillion data points in seconds. Ask it a month from now, and you may talk to an updated version of the form partially modified with the research and development conducted by the original model. This is not theoretically – it is already happening on margins and acceleration.

The effects of American national security, economic competitiveness and civil society may be vast. If these models can learn from their mistakes, independently adapt, and eventually design and train their successors, then whoever controls this process. a lot.

That is why we are in Congress, especially those who are working on technology, defense and foreign policy, we must start to believe with this issue and behavior now, and not yet it is too late.

The Chinese Communist Party has already realized the strategic value of the automatic spontaneity organization. In 2017, the “new artificial intelligence development plan”, which aims to control the development of global artificial intelligence by 2030. This plan is not only related to research financing, it is a belt initiative and a way for Amnesty International: exporting Chinese influence, including party friendly systems abroad, and erosion of American technological leadership from within allied economies.

If the United States fails to lead in the responsible development of automatic artificial intelligence systems, we risk more than economic decline. We risk controlling the waiver of a future formed by the algorithms of black boxes and self -directed machines, some of which are not in line with democratic values ​​or basic human safety.

The United States does not need to imitate China’s approach from top to bottom. Our strength has always come from private innovation, scientific openness, free market, and deep respect for freedom. We must find a way to encourage research and progress in ​Research and automatic development capabilities. At the same time, we must invest in decisive areas to oversee the research and development of automatic artificial intelligence, such as the systems of interpretation and control of artificial intelligence, which were identified in President Trump’s plan of artificial intelligence.

Ensure that the United States remains prominent in ​to​ Amnesty International ​development​​ Without losing the vision of transparency, accountability and human control​ The correct questions are required now:

• When does the self -improvement of the artificial intelligence system express a threshold that requires organizational care?
• What are the frameworks, or must be built, to ensure human control and development systems are increasingly artificial intelligence?
• How do we evaluate artificial intelligence systems and verify themselves healthy products?
• What are the mechanisms necessary for Congress to stay appropriately with the research and automatic development that occurs within private artificial intelligence companies?
• How can Congress enhance innovation with protection from misuse or weapon of these technologies?

I do not pretend to have the final answers. But I believe a firm belief that the pace and depth of this discussion (and the resulting work) must be accelerated and intense, before the next generation of artificial intelligence systems begins writing the future without us in the episode. AI is already progressing greatly, driven by computing computing, data and algorithms; Since automatic systems create artificial data and discover new algorithms, the progress of artificial intelligence may turn from si to the explosive.

This is not an invitation to regulate a sweeping, nor is it not an invitation to alarm. It is an invitation to avoid sleeping on controls.

ATM research and development will be a distinctive feature of global competition in the coming years. The United States must make sure that, not our opponents, we set the moral and strategic boundaries of this technology. This work begins here, in Congress halls.

Nathaniel Moran is the first boycott of Texas in Congress.