Quality issues with training data are holding back AI projects
For many organizations, AI and machine learning are seen as a route to greater efficiency and competitive advantage.
But according to a new study conducted by Dimensional Research for Alegion almost eight out of 10 enterprise organizations currently engaged in AI and ML report that projects have stalled, and 96 percent of these companies have run into problems with data quality, data labeling required to train AI, and building model confidence.
"The single largest obstacle to implementing machine learning models into production is the volume and quality of the training data," says Nathaniel Gates, CEO and co-founder of Alegion, a training data platform for AI and ML initiatives. "This research reinforces our own experience, that data science teams new to building ROI-driven systems try to tackle training data preparation in house, and get overwhelmed."
The study finds 70 percent report that their first AI/ML investment was within last 24 months and over half of enterprises report they have undertaken fewer than four AI and ML projects. Just half of enterprises have released AI/ML projects into production.
With AI becoming a growing enterprise priority, data science teams are under more pressure to deliver projects but frequently are challenged to produce training data at the required scale and quality. 78 percent of their AI/ML projects stall at some stage before deployment and 81 percent admit the process of training AI with data is more difficult than they expected.
76 percent combat this challenge by attempting to label and annotate training data on their own, while 63 percent go so far as to try to build their own labeling and annotation automation technology. 71 percent of teams report that they ultimately outsource training data and other ML project activities.
The full survey is available from the Alegion site.