Are AI Governance Programmes Needed for Compliance with AI Legislation?
5th April 2024
Once enforcement of the new AI act begins, organisations may face fines as large as €35 million as retribution for using forbidden AI tools. However, Gartner has suggested that such punishments can be avoided if AI governance programmes are implemented within the organisation. An AI governance programme can help organisations catalogue and categorise their AI tools and systems, meaning they can identify and consequently address any prohibited AI, to ensure it is safe and ethical.
Tendü Yoğurtçu, CTO at Precisely, agrees with Gartner, and believes AI governance is essential for any organisation that wishes to protect themselves from penalties, whilst still getting the most out of AI:
“As the utilisation of artificial intelligence (AI) increases rapidly, lawmakers and regulators worldwide are paying close attention – with the EU AI Act approved by EU parliament only last month and news of a landmark AI safety agreement signed by the US and the UK. It seems clear that organisations must have the correct guardrails in place to protect their business – while still allowing room for innovation. To achieve this, establishing a robust data integrity strategy, including effective data and AI governance, is essential.
“A decade ago, data governance was fundamentally a technical undertaking centred around compliance. These functions were often performed by the IT department, and the primary purpose was to improve the quality of internal data for use by specifically identified teams. With the accelerated digital transformation and adoption of AI, data has become the most vital corporate asset, and, alongside that, data governance must become an enterprise-wide priority.
“A strong data governance framework allows organisations to easily find, understand, and leverage critical data – leading to more accurate and informed decisions and reporting. It provides a crucial understanding of the meaning, lineage, and impact of data, allowing businesses to stay ahead of changing regulatory landscapes, while ensuring that AI models are fuelled with trustworthy data for outcomes that can be relied on.
“Therefore, now that more companies are using AI to analyse, transform, and even produce data, a strong governance framework for AI along with the governance of the data is paramount. From ensuring quality and compliance of generative AI output to tracking the lineage of how these decisions are made, transparency of automation is critical for ensuring compliance and competitiveness.”