China leads in genAI adoption, business leaders confident in skills: report

China is currently leading the pack in adoption of generative AI, with 83% of Chinese business decision makers reporting that their organisation uses the technology according to a July survey.

Of the 17 countries represented in the survey, carried out by SAS in collaboration with Coleman Parkes, the UK came in second place with 70% reporting genAI usage, followed by the United States with 65%, ahead of the global average of 54%.

With China’s Xinhua News Agency recently reporting that more than 600 million people in mainland China are using Large Language Models (LLMs), and the Chinese government giving the green light to an average of more than one AI model per day for the past six months, the report puts a spotlight on the nuances of genAI use in the world’s second-largest economy, as well as how it measures up to other parts of the globe.

Chinese business leaders are confident in genAI tools and expertise

Respondents from organisations in China feel confident that they have the necessarily tools and know-how to not only invest in generative AI but also to implement it.

When asked about implementation challenges, Chinese respondents were far less likely than the global average to cite a lack of internal expertise or appropriate tools: only 31% of respondents said they lacked the right tools to implement generative AI (compared with 47% of respondents globally), while just 21% said they lacked internal expertise (39% globally).

Marinela Profi, Strategic AI Advisor at SAS, commented in the report that, “LLMs alone do not solve business problems. GenAI is nothing more than a feature that can augment your existing processes, but you need tools that enable their integration, governance and orchestration. And most importantly, you need people that can use tools to ensure the appropriate level of orchestration.”

Among all respondents to the survey, data privacy and data security ranked as the two top concerns in implementing genAI, cited by 76% and 75% respectively. However, more than half of respondents (51%) expressed concerns about the need for in-house talent and skills.

C-suite urged to clarify AI productivity expectations and focus on skill building

GAI training was found to be particularly lacking in one specific area: governance and monitoring. According to SAS, fewer than one in 10 respondents (7%) reported a “high” level of governance and monitoring training for GenAI. Thirty-two percent reported an “adequate” level, while 58% – a clear majority – said their governance and monitoring training was “minimal”.

Adding to this, when asked about their organisational governance frameworks for generative AI, just 5% of respondents (among those who were using, or planning to use, generative AI) said they had a “well-established and comprehensive” governance framework. More than half (55%) said their governance framework was “in development”, while 28% described it as “ad hoc or informal”. Roughly 1 in 10 (11%) called their GenAI governance framework “nonexistent”.

APAC and China lead on organisational usage policies, regulatory compliance

One fundamental layer of governance that a majority of organisations do have in place is an organisational usage policy. Sixty-one percent of respondents said that their organisation had “a GenAI policy that dictates how employees can and cannot use it for the business” – unquestionably a fundamental layer in any business’s implementation of genAI.

The Asia-Pacific region (comprising China, Australia, Japan and the UAE/Saudia Arabia) led other global regions from the survey in implementing this type of policy: a full 71% of APAC organisations said that they have an organisational usage policy in place. The region with the next-highest level of preparation in this area was North America, with 63%.

APAC also led the way in compliance with genAI regulations – an emerging, but rapidly-moving, area of law. Seventy percent of APAC respondents said that they were either fully (13%) or moderately (57%) prepared to comply with current and upcoming regulations concerning generative AI, ahead of a global average of 58% (10% of whom are fully prepared, while 48% are moderately prepared).

SAS’ survey compared how well-prepared organisations are to comply with genAI regulations across regions. (Image: SAS & Coleman Parkes)

China reported the highest level of full preparation for genAI regulations, at 19%, followed by Australia (15%) and the United States (14%). This is unsurprising when considering China’s regulation-forward approach to generative AI: Matt Sheehan, a fellow at the Carnegie Endowment for International Peace, has credited China with “rolling out some of the world’s earliest and most detailed regulations governing artificial intelligence”.

China hasn’t hesitated to put interim measures in place to regulate generative AI while fuller guidelines are developed, and all publicly-released LLMs require government approval.

Last month, China also published the Shanghai Declaration on Global AI Governance at the World Artificial Intelligence Conference: a set of governance guidelines and commitments for AI that it has called on other countries to adopt in order to establish international co-operation around generative AI.

China’s implementation of genAI: widespread, but still largely experimental

While China’s adoption of generative AI is undeniably widespread, Chinese organisations haven’t yet implemented genAI as fully as they might.

When SAS surveyed respondents about the extent to which their organisation is using generative AI, 19% of China’s organisations said that they are “using generative AI and have fully implemented it (i.e., integrated GenAI into our regular processes)” – ahead of the global average of 11%, but behind the world leader in full implementation, the US, at 24%.

Meanwhile, 64% of respondents from China said that their organisation is “using generative AI but haven’t yet fully implemented it (i.e., running initial tests/experiments)” – well above the global average of 43%. (The UK ranked second in genAI experimentation with 58%).

Given China’s emphasis on careful regulation and authorised approval of generative AI, it makes sense that many organisations would be conducting initial tests prior to fully integrating genAI within their processes. It’s clear that China is all-in on generative AI, but China’s organisations are pacing themselves even as they collectively embrace this new technology.