I left the academic AI research after my graduation from the university of Illinois in 2008.
To apply my engineering skills to concrete business problems, I became a developer and then an enterprise architect. With this background, I see with great interest the current rising from the lab AI to the public applications.
In 2016 US government under the Obama launched an initiative to draw a strategy on AI.
In 2018 China ministry of information technology & France government issue a report on AI.
This article compares the topics and conclusions from those reports. For more details the reports are publicly available on the internet:
· Chinese 2018 report on IA: http://www.miit.gov.cn/n1146295/n1652858/n1652930/n3757016/c5960820/content.html
· French 2018 report: https://www.aiforhumanity.fr/pdfs/9782111457089_Rapport_Villani_accessible.pdf
First, all the 3 reports share a common assessment of the situation and propose various actions for the government to encourage the progress and integration of the technology to the economy.
In common, the 3 governments are aware of the disruption power of AI technologies to all sectors. In deed AI, capabilities improve computer with
· human-like understanding on text, voice, image, video,
· prediction learned from historic data,
· autonomous intelligence to execute specific task.
The availability of data and powerful hardware fueled the machine learning algorithms with performance near or above human capabilities on certain tasks. For example, some AI application beat human accuracy on cancer diagnosis based on medical imaging.
The business opportunities for automation are tremendous especially on government-controlled sectors such as: education, transportation, healthcare, defense.
To achieve AI integration in the economy, they have identified as top priorities
· To encourage research to advance improve the performance and the reliability of AI algorithms.
· To accelerate cycle to transfer from R&D advances to the industry products & services.
The 3 governments have identified as solutions:
· To collaborate openly at the research level to the advance of the AI field,
· To address the shortage of the IA skills by training, attracting and retaining talents AI researchers, AI software engineers,
· To build of sectors-based AI platforms to collect open dataset, standardize data, experiment AI applications, test and produce applications.
Let’s focus on each country specific plan,
China has the most objective oriented plan. By 2020 AI should be part of the economy with augmentation of existing manufacturing capabilities.
Clear quantitative objective (specified technically) have been made on AI products, hardware and organizations.
· AI products include Robot, autonomous ground and aerial products, medical imaging diagnosis system, Video image recognition, intelligent voice, intelligent translation, smart home device, security products,
· AI infrastructure include Neuromorphic chip optimized in performance and energy consumption, AI sensor with open source development kit.
The US 2016 reports collect AI experts and humanities (social, legal) under 4 public workshops.
The economic approach is to foster public research in open source and rely on private company to develop innovative AI products.
· But the report highlights also the risks to manage for the future:
· Anticipate the need to retrain human in value added job with collaboration of AI application,
Through automation AI will replace human on specific tasks.
· Control the bias of the AI application, the AI can replicate human bias opinion,
· Develop explicability of the algorithms, today AI works well but it is a black box,
· Regulate the use of autonomous weapons,
· Mitigate risks on cyberwarfare,
The reports conclude on a series of recommendation for the government.
France report in 2018 was led by the mathematician and member of parliament Cedric Villani.
The report follows the topics addressed in the US report and ends up with similar conclusions contextualized with the European situation.
The French state can play an incentive role to order AI products, animate, promote sector innovations.
For the R&D, the report recommends to foster in France the research with IA dedicated institute 3IA and to form with European academic labs cluster for research in IA. Discovery in AI should be shared openly.
On the workforce, the report recommend to retrain on skills where human outperform AI such as : general intelligence, problem resolution , creativity, sociability.
Specifically, the report addresses opportunities of AI to tackles climate change problems and the call for an ethicial AI. For the Ethics, at the European level, citizen privacy and security is highly regarded. The recent in Mai 2018 application of the GDPR constraints for example companies to manage personal data with citizen consent.
To conclude, US, China and France acknowledge the business opportunities of the AI technologies.
All the 3 countries consider “human brain” capabilities in IA
· Researchers to create new algorithms and,
· Engineers to decline into economically viable product as soon as possible.
China has set ambitious goals for 2020 to integrate into its economy.
US and France emphasis the risks and social challenges on workforce and ethics for the society.
As of today, it is uncertain if the US and French government will transform the recommendation into a strategic plan.