Watson-全球人工智能采用指数2021（英文）-2021.5-13页_4mb

报告摘要

2021 IBM Global AI Adoption Index Summary

Core Content

The IBM Global AI Adoption Index 2021, conducted in partnership with Morning Consult, provides insights into the global state of AI adoption across various industries and regions. The survey included 5,501 IT professionals from 15 markets, including China, France, Germany, India, Italy, Latin America, Singapore, Spain, the United Kingdom, and the United States. The research was conducted in April 2021 via an online platform, targeting individuals with significant input into their firm's IT decision-making.

AI Adoption Overview

Almost one-third (33%) of IT professionals reported that their company is using AI.
Almost half (47%) are exploring AI, showing a continued growth in interest.
43% of businesses accelerated AI adoption due to the COVID-19 pandemic, especially in larger organizations.
74% of companies are exploring or deploying AI, indicating a strong trend in AI integration.
34% of global IT professionals are in the phase of analyzing data to build and scale AI but have not rolled it out yet.

Key Drivers of AI Adoption

The top three reasons for AI adoption are:

Advances in AI that make it more accessible (46%)
Business needs (46%)
Changing business needs due to the pandemic (44%)

Barriers to AI Adoption

The top three barriers to AI adoption are:

Limited AI expertise or knowledge (39%)
Increasing data complexity and data silos (32%)
Lack of tools or platforms for developing AI models (28%)

Data complexity and silos are more critical for larger companies.
Lack of AI skills is the main barrier for smaller firms.

AI Investment and Use Case Trends

43% of companies accelerated AI adoption due to the pandemic.
31% of larger companies reported a faster rollout compared to smaller ones.
38% of IT professionals cited improving employee productivity as a key factor in AI adoption.
36% reported that better customer interaction influenced AI use.

Top AI Investment Areas (Next 12 Months)

Area	%
Data security	31%
Automation of processes	25%
Customer care	25%
Virtual assistants/smart chatbots	20%
Business process optimization	19%
Fraud detection	16%
Sensor data analysis (IoT)	15%
AI monitoring and governance	14%
Marketing	14%
Supply chain	11%
Personal security	11%
Predictive decision making	10%
Image recognition	9%
Natural language processing (NLP)	7%
Search	7%
Recommendations	6%
Healthcare diagnostics	6%

Trustworthy AI

84% of IT professionals believe the ability to explain AI decisions is important.
90% of companies using AI consider this ability critical for maintaining brand integrity and customer trust.
Over three-quarters (76%) of IT professionals believe it is critical to trust AI's output as fair, safe, and reliable.
86% of respondents agree that consumers prefer companies with transparency and ethical frameworks.

Biggest Barriers to Developing Trustworthy AI

Barrier	%
Lack of skills or training	65%
AI governance and management tools that don’t work across all data environments	62%
AI outcomes that are not explainable	58%
Lack of regulatory guidance	58%
Lack of company guidelines for ethical AI	58%
Building models on biased data	58%

NLP Adoption

Almost half (49%) of respondents use NLP, and 25% plan to use it in the next 12 months.
Over half (52%) use or plan to use NLP to improve customer experience, and 43% to increase cost efficiency.
Top NLP use cases include:
- Email or text classification (35%)
- Machine translation (34%)
- Virtual agents for customer service (34%)
- Call center automation (33%)
- Survey analysis (31%)

Automation Trends

80% of companies are using or planning to use automation tools.
Top reasons for automation use:
1. Driving greater efficiencies (58%)
2. Saving costs (58%)
3. Giving valuable time back to employees (42%)
38% of companies used automation to improve employee productivity during the pandemic.
China has the highest adoption rate of automation (92% using or planning to use).
UK has the lowest adoption (32% not using, 49% already using, 19% planning to use).

Popular Automation Use Cases

Use Case	% Using	% Interested in Using	Total
Network performance	56%	32%	88%
Integration of apps and data	54%	35%	89%
Business process management (BPM)	45%	40%	85%
Application performance management (APM)	40%	43%	83%
Observability	38%	43%	81%
Process and task mining	37%	43%	80%
Robotic process automation (RPA)	33%	44%	77%

Automating IT operations is the top use case for automation.
Smaller companies focus more on activity monitoring, while larger companies emphasize IT operations.

Data Access and Management

87% of IT professionals believe it is very or somewhat important to be able to build and run AI projects anywhere data resides.
72% of IT professionals in India are confident in their data access tools, compared to 23% in China and 44% in the US.
67% of global IT professionals draw from more than 20 data sources for AI, BI, and analytics.
Larger companies are more likely to use multiple data sources (75% vs. 34% for smaller companies).

Methodology

Conducted online via Morning Consult's proprietary network in April 2021.
Sample included IT professionals with significant input into their firm's IT decision-making.
Represented a mix of small and large firms, with 28% from firms with more than 1,000 employees and 25% from smaller businesses (50 employees or less).
VP level or above (25%) and directors/senior managers (75%) were included in the sample.

Summary of Key Insights

AI adoption is growing globally, driven by pandemic impacts, business needs, and technological advances.
Data complexity, lack of skills, and tool limitations are the main barriers to AI adoption.
Automation is becoming a core part of business operations, with a focus on efficiency, cost savings, and employee productivity.
NLP is widely used and considered for future growth, especially in customer service and experience enhancement.
Trust in AI is a critical factor, with a strong emphasis on explainability, fairness, and ethical use.
Data accessibility and integration remain significant challenges, particularly in multi-cloud and multi-vendor environments.