A HPenterpriseA new report from Hewlett Packard Enterprise (HPE) shows that companies eager to implement artificial intelligence are struggling with the processes needed to effectively deploy models. The report, titled "Building an AI Advantage," surveyed more than 2,400 IT leaders from 14 countries.
Source Note: The image is generated by AI, and the image is authorized by Midjourney
Respondents were from organizations with more than 500 employees, covering a variety of industries including healthcare, manufacturing and financial services. The report found that companies looking to implement AI have difficulty with the basic process of preparing data for AI models. The survey results showed that only 7% of the 2,400 IT professionals surveyed were currently able to run data synchronization in real time, and only 26% were able to runadvancedAnalytical applications: Less than 60% respondents say their organizations currently have the capabilities to access, store, and recover data, which can slow down the model development process.
The report warns that businesses that fail to address these implementation issues risk having their models produce inaccurate outputs. The report also found other issues, including a failure to understand the level of compute and network required to run their applications. IT leaders surveyed expressed confidence in the outlook, with 93% saying their network infrastructure was appropriate and 84% saying their compute storage was adequate. However, the report found that less than half of respondents admitted that they did not fully understand the level of AI workloads required in terms of training and inference.
There is also a policy issue, with more than a quarter (28%) of respondents describing their company's approach to AI as "fragmented." The report also found that about 35% of IT leaders said their company has created separate AI strategies for each function, while 32% said their employer has developed completely different plans. In addition, HPE also found that there is a lack of consideration of ethical issues during deployment. About a quarter (22%) of respondents admitted that their company's legal team was not involved when developing an AI strategy.
Eng Lim Goh, senior vice president of data and artificial intelligence at HP Enterprise, said: "Enterprises must carefully weigh the balance between being early movers and not fully understanding the gaps at each stage of the AI lifecycle, otherwise large-scale capital investments may result in negative returns." Overall, the report reveals the challenges faced by enterprises in deploying AI models, warning that enterprises need a more comprehensive approach to promote the comprehensive deployment and application of AI.