Converseon reposted this
the eu ai act finally becomes law on august 1. as has been cited elsewhere, for individuals or companies that are deploying or implementing AI systems (referred to as "deployers" under the Act), here are three key obligations to keep in mind: 🔥 Implement a risk assessment system; 🔥 Adopt transparency principles and processes; and 🔥 Design responsible AI governance. While gen AI is a focus it does not discriminate in other uses of ai including data classification and predictive ai. for those in social listening and media evaluation, market research, consumer intelligence using these technologies, remember you are responsible even if you did not build the model. i’ve been a bit repetitive on this lately but there are implications - Commit to data quality - Know your model accuracy! - embrace model observability, transparency and validation - ensure human in the loop governance. Not sure where to start? we can help. our Conversus NLP platform provides the full range of “trusted” ai features needed to abide by these standards. just dm me. the good news? doing so will also drive more business value, adoption, and trust in your data and insights. (hat tip to gartner too for their strong coverage of this framework) -
Kathleen Fetters
1mo
As a social intelligence practitioner for over 15 years, I consider data validation a standard protocol that too often is skipped for one reason or another. So I say keep being "repetitive" Rob Key!
This is a highly insightful breakdown of the obligations under the upcoming EU AI Act. Ensuring data quality and accuracy, along with transparency and human governance, will undeniably be crucial for compliance and trust-building. What do you believe is the biggest challenge deployers might face in aligning with these new regulations? We’re curious to see how different sectors adapt their AI strategies in response.