Speaking at the World Economic Forum a few days ago, Liz Centoni explained a wide-angle view that it's about the data that feeds AI models.
Data and context to customize AI models derives distinction, and AI needs large amounts of quality data to produce accurate, reliable, insightful output.
Some of the work that's needed to make data trustworthy includes cataloging, cleaning, normalizing, and securing it.
It's underway, and AI is making it easier to unlock vast data potential.
AI has and will continue to be front-page news in the year to come, and that means data will also be in the spotlight.
Data is the backbone and the differentiator for AI, and it is also the area where readiness is the weakest.
The AI Readiness Index reveals that 81% of all organizations claim some degree of siloed or fragmented data.
While siloed data has long been understood as a barrier to information sharing, collaboration, and holistic insight and decision making in the enterprise, the AI quotient adds a new dimension.
With the rise in data complexity, it can be difficult to coordinate workflows and enable better synchronization and efficiency.
Leveraging data across silos will require data lineage tracking so that only the approved and relevant data is used, and AI model output can be explained and tracked to training data.
We'll look back a year from now and see 2024 as the beginning of the end of data silos.
Emerging regulations and harmonization of rules on fair access to and use of data, such as the EU Data Act which becomes fully applicable next year, are the beginning of another facet of the AI revolution that will pick up steam this year.
Unlocking massive economic potential and significantly contributing to a new market for data itself, these mandates will benefit both ordinary citizens and businesses who will access and reuse the data generated by their usage of products and services.
According to the World Economic Forum, the amount of data generated globally in 2025 is predicted to be 463 exabytes per day, every day.
The sheer amount of business-critical data being created around the world is outpacing our ability to process it.
It may seem counterintuitive that as AI systems continue to consume more and more data, available public data will soon hit a ceiling and high-quality language data will likely be exhausted by 2026 according to some estimates.
Both private and synthetic data, as with any data that is not validated, can also lead to bias in AI systems.
Misuse of private data can have serious consequences such as identity theft, financial loss, and reputation damage.
Organizations must ensure they have data governance policies, procedures, and guidelines in place, aligned with AI responsibility frameworks, to guard against these threats.
Recognizing the urgency that AI brings to the equation, the processes and structures that facilitate data sharing among companies, society, and the public sector will be under intense scrutiny.
This Cyber News was published on feedpress.me. Publication date: Fri, 26 Jan 2024 13:13:04 +0000