Transformative release couples deeper integration with OpenAI’s GPT models with stronger process and governance features to help enterprises drive results with AI while reducing risk
Dataiku, the platform for Everyday AI, has unveiled Dataiku 12, their latest breakthrough release packed with transformative new features designed to help organisations confidently harness AI in their advanced analytics initiatives with enhanced transparency and strengthened governance features.
As businesses strive to stay ahead of advanced analytics and AI innovation, ChatGPT and Large Language Models have become central to boardroom discussions. Companies must balance the pressure to deploy AI for a competitive edge with managing potential risks to their reputation and operations. For many, the risks are overwhelming due to poor visibility and inconsistent processes across data, analytics, and AI tools used by various teams.
Dataiku 12 addresses these challenges while providing total confidence, control, and trust in AI outputs. The new features include:
- OpenAI GPT Integration: Dataiku allows business users to incorporate OpenAI’s GPT models into data projects by extending datasets and performing tasks using a visual interface and natural language prompts, all while maintaining transparency and trust in project outputs.
- Causal Machine Learning (ML): Causal predictions ensure that correlation is not confused with causation when doing so would be harmful to the business or other stakeholders. Dataiku democratises these capabilities, ensuring that anyone building ML models can understand the “why” behind their results.
- Universal Feature Importance: Some ML models offer limited decision-making explanations, undermining stakeholder trust. Dataiku centralises model explainability, so teams have a consistent way to explain models that build confidence and trust with business users.
- Model Overrides: A core principle of AI safety is maintaining human oversight. In some cases, predictive models don’t have the best or safest answer, so Dataiku allows experts to create strict rules to enforce model outcomes for known cases based on real-world experience.
- Model Risk Project Views: With AI projects across teams at different stages of development, it is difficult to manage risks and allocate resources. Dataiku allows business and analytics leaders to easily spot and mitigate risks in AI projects, increasing trust in project outputs.
- Transparent Automated Feature Generation: Feature engineering can be a “black box” in AI modelling, increasing perceived risk. Dataiku gives people transparency and control over this ML technique, helping to understand where new features originate.
“As we advance into the era of AI, delivering trust via transparency and governance is now essential for any business aiming to stay at the forefront of innovation. Dataiku 12 directly addresses this need,” said Florian Douetteau, co-founder and CEO at Dataiku. “We’re empowering companies to confidently explore the vast potential of AI while providing robust mechanisms for oversight and risk management. This fusion of power and control is the key to unlocking the full potential of AI in the most responsible way possible.”