Appen Limited, a global AI leader in providing data sources, data preparation and human model evaluation at scale, has released its highly anticipated annual report “AI status and Machine learning report.”

The State of AI and Machine Learning report is an annual report focused on the strategies implemented by companies of all sizes across all industries as they deepen their AI maturity . The latest edition is the eighth published by Appen, and it highlights key approaches to data management and security, responsible AI, and external data providers and their role in driving progress.

Main findings of the report

The main findings of the report related to supply, quality, evaluation, adoption and ethics.

A key finding of the report is that 51% of respondents agree that data accuracy is critical to their AI use case. It’s well known that high-quality, accurate data is critical to the success of AI models, but many business leaders have a significant gap between the ideal and reality in achieving data accuracy, according to the report.

Another key takeaway is that companies are increasingly turning to responsible AI and maturing their strategies. A growing number of business leaders and technologists are working to improve the quality of data that powers AI projects, which promotes inclusive data sets and unbiased models. The report found that 80% of respondents think data diversity is “extremely important” or “very important.” He also found that 95% of respondents agree that synthetic data will be a key player in creating inclusive datasets.

Mark Brayan is CEO of Appen.

“This year’s State of AI report reveals that 93% of respondents believe responsible AI is the foundation of all AI projects,” Brayan said. “The problem is that many face the challenges of trying to build great AI with poor datasets, and that creates a significant barrier to achieving their goals.”

Here are some of the other key points from the report:

  • Supply: 42% of technologists say the data sourcing stage of the AI ​​lifecycle is very difficult, and business leaders were less likely to report data sourcing as very difficult (24%).
  • Quality: More than half of respondents say data accuracy is critical to AI success, but only 6% said they had achieved data accuracy above 90%.
  • Evaluation: There is a strong consensus around the importance of human machine learning in the loop, with 81% saying it is very or extremely important. 97% said human feedback in the loop is important for accurate model performance.
  • Adoption: Technologists are divided as to whether their organization is ahead or even with others in their industry. US respondents are more likely to say their organizations are ahead of others in their industry when it comes to AI adoption compared to European respondents.
  • Ethics: 93% of respondents agree that responsible AI is the foundation of all AI projects within their organization.

Sujatha Sagiraju is a Product Manager at Appen.

“The majority of AI efforts are spent on data management for the AI ​​lifecycle, which means it’s an amazing business for AI prospects to manage alone – and that’s the field with which many are struggling with,” Sagiraju said. “Sourcing high-quality data is critical to the success of AI solutions, and we’re seeing organizations highlight the importance of data accuracy.”

Wilson Pang is CTO at Appen.

“Data accuracy is critical to the success of AI and ML models because qualitatively rich data yields better model results and consistent processing and decision-making,” Pang said. “For good results, datasets must be accurate, complete and scalable.”

You can find the full report on the state of AI and machine learning here.