Plenty has been written about data and big data in the past couple of years but relatively little about the analysis, interpretation and the productive use of all this data. Historically, it has been prohibitively expensive to collect, tag and organise datasets into a format required for real time analysis. Now, for the first time, artificial intelligence (AI) is enabling the processing of unstructured data sets on a huge scale.
Image recognition was arguably the most impactful first-wave application of AI technology. The medical sector quickly looked to adopt this technology thanks to the relatively simple machine learning models required and a ready availability of structured, image data sets that can be used for training algorithms.
Looking forwards, natural language processing (NLP) is the second-wave application. It is one of the most challenging AI applications as massive numbers of parameters are required in the NLP models to enable the knowledge transfer among different tasks during the training process. This increases computing intensity dramatically though the potential applications derived from NLP models would be truly transformational for many industries.
Many of you will already have come across increasingly fluent chatbots and digital assistants that can change how companies interact with existing or potential customers. E-commerce recommendation engines are becoming more accurate and the human/digital interface less frictional. Outside the technology sector, in industries undergoing more significant technological and digital transformations, we are already seeing early examples of NLP experimentation and adoption in legal, financial and insurance settings, very much driven by the most forward-looking companies.
AI development is ongoing and, looking ahead, the benefits of these technologies will continue to move beyond the technology sector.Developments such as these are not linear. For example, recent political and regulatory attention has been focused on the vast data sets that are necessary to train the neural networks, the core framework of AI. Any reduction in the availability and/or accessibility of these data sets would be detrimental to the sector’s development. While there has rightly been increased scrutiny on the use of data that includes sensitive or personally identifiable information, the impact on overall AI development has not been noticeable.
Then there is the impact of coronavirus. In the near term, any company that has China exposure either directly or indirectly will be impacted by coronavirus-related slowdowns due to the disruption in local supply chains. The spread of the virus beyond China has raised concerns the global economy may experience a slowdown due to further disruption.
AI development is ongoing and, looking ahead, the benefits of these technologies will continue to move beyond the technology sector. As the adoption of AI grows and the analysis of data it allows improves, we expect to see companies evolve and adapt to the first two waves as well as whatever any third wave of applications may bring.