Government organisations are historically slow to adopt new technology – but there is a huge opportunity with artificial intelligence (AI) that they should not miss out on. In the private sector, it has already transformed how we work, streamlining manual and tedious tasks, enhancing employee skills and fostering more rapid innovation. But without proper data foundation and security measures, there are inherent risks. Here are three ways government entities should prepare for safe and effective AI usage.
1. Prepare data
It is vital to must organise and centralise data in order to improve data quality and enhance AI recommendations. While many organisations have their data in the cloud now (after years of on-prem storage) we’ve found that many house data in multiple clouds and collaboration systems. In fact, in AvePoint’s inaugural AI and Information Management Study, which surveyed over 750 digital workplace decision makers globally, we asked respondents to select all the places they store data, and the results were unsurprising, but illuminated a key challenge. Eighty-seven percent of survey respondents selected the cloud, but 51 percent also had data in self-hosted storage and 46 percent had it in physical documents.
Because AI algorithms need access to all relevant information to build appropriate learning models, it’s critical to coalesce data so that the AI algorithms are pulling from one location. Otherwise, recommendations will be incomplete or inaccurate.
2. Secure data
Data security is extremely important to prioritise and you have to get it right when implementing AI. GDPR means that European organisations are acutely aware of data security and privacy. However, AI implementation provides new considerations and requires specific AI training on safe usage. That’s why it’s necessary to understand where sensitive data and overshared content lives in digital work environments and implement robust security measures to secure that data.
Similarly, investing in employee training so that users know what type of information can be shared when using public AI models and what should only be accessed within the organisation’s secure
tenant is a need not a want. Otherwise, material non-public information or sensitive citizen data may be inadvertently shared. And this is already a key issue: according to Gartner, among organisations who have faced an AI security or privacy incident, 60 percent reported data compromise by an internal party.
3. Optimise operations
Finally, data management and governance controls are critical to fuelling sustainable AI success. Information management (IM) strategies – which include archiving and retention policies, lifecycle management, solutions, automated data classification, and data analysis and discovery – help with this by ensuring which data should be retained, archived or pruned. This not only improves AI output but can also reduce storage costs.
As AI comes to full swing, and with Gartner predicting 10 percent of all data generated will come from AI by 2025, organisations will need to focus even more on information management to ensure their data is of high quality and trustworthy. Without these solutions, they are setting themselves up for failure.
Ultimately AI can be a powerful solution especially for government entities looking to drive real change. But with great power comes great responsibility. That’s why it’s key to effectively prepare and secure data and optimise operations when implementing and using AI.
