AI systems' ability to be transformative and novel will be the primary evaluation criteria in 2024, according to predictions from Steve McMillan, President and Chief Executive Officer at Teradata. Following 2023’s surge in interest around large language models (LLM) and ChatGPT, the focus will be on AI driving creative breakthroughs and industry transformations in 2024. McMillan anticipates significant investments in generative AI initiatives that offer significant improvements, over mere novelty, in outcomes. Developing solutions that augment the efficiency and competence of internal operations will stand vital, leading to embedded AI in organisational products and services.
"Getting the data right" will be indispensable as businesses integrate AI-driven automation and AI, together with machine learning (ML), into virtually all aspects of their decision-making processes. McMillan emphasised that for AI projects, the underlying data must be clean, accurate, and dependable.
2024 will also witness an increment in the demand for enterprise-scale AI experimentation and discovery as more organisations seek to liberate data, particularly open data. The evolving solutions incorporating serverless query engines, integrated data with advanced LLMs, and accessible open table formats will optimise experimentation needs. McMillan believes these solutions will simplify the process for data scientists, engineers, and developers to explore and unearth innovative use cases on-demand and at scale.
Trust, ethics, and sustainability will also dominate the discourse in 2024 AI-related matters. The emphasis will be on trusting information and ensuring transparent datasets. Without this transparency, comfort with the outcomes becomes doubtful as the quality of predictions or insights directly correlates with the quality of informing data.
Steve McMillan opined that the inception point of future AI ventures should ascertain the trustworthiness of the data asking whether it is clean and reliable. There will also see a heightened emphasis on data ethics as governments and consumers alike demand increased transparency and accountability within AI, particularly concerning data used in training AI systems. Additionally, as businesses will undergo thorough scrutiny on the immense energy used by AI projects, we can perhaps anticipate a rise in small to medium language models that are not only customisable, accurate, and secure but also more efficient.
"In 2024. It will become increasingly apparent that AI must be trusted, ethical and sustainable. Everyone will be talking about trusted information and trust in information," McMillan says.
"This is important because without knowing and trusting the data sets used for AI projects, can you really be comfortable with the outputs? The quality of predictions or insights is only as good as the data that informs it."