There have been predictions that the Trump administration won’t only expand the export ban but in addition ban cloud use and even access to large language models (LLM) that exceed certain parameters in an effort to stay ahead of the synthetic intelligence (AI) competition with China.
IEEE Spectrum, the technology media of the Institute of Electrical and Electronics Engineers (IEEE), made a prediction on the twenty seventh (local time) about how the newly appointed administration will affect the technology sector.
Areas where the most important changes are expected include AI, home appliances, cryptocurrency, energy, and transportation.
Amongst them, regarding AI, it was expected that the Trump administration would encourage the usage of AI for national security. He also identified that there’s a high possibility of expanding export restrictions related to AI technology to remain ahead of China.
Particularly, he identified that although currently American semiconductor corporations cannot sell cutting-edge chips to Chinese corporations, there is a big loophole on this rule. All you have got to do is enroll for a US-based cloud computing service.
This problem has been identified by the Biden administration since last yr, and reports have emerged that giant cloud corporations are providing supercomputing infrastructure to Chinese corporations through expedient means. It seems that the Trump administration won’t just sit by and watch this.
They said export controls could even be expanded to limit Chinese corporations’ access to the U.S. foundation model based on parameters. On this case, China’s access to the very best performing open source model have to be blocked.
In reality, early this month, it was reported that China had developed a military chatbot using Meta’s ‘Rama’. As well as, there have been criticisms inside China that China is in truth depending on American technology, including open source.
IEEE Spectrum believes that this point may even be subject to regulation.
“President-elect Trump pledged through the election campaign to rescind AI executive orders enacted by the Biden administration, but organizations just like the AI Safety Institute will survive,” said Doug Calidas, senior vp for presidency affairs on the American Responsible Innovation advocacy group. It was predicted.
This organization, which is collaborating with OpenAI and Antropic to check the security of Frontier models before launching them, is already receiving sufficient support from each parties. Whether it is eliminated blindly, there will probably be a burden of getting to create an analogous organization again.
Reporter Lim Da-jun ydj@aitimes.com
