Back to blog home
Responsible AI Shifts Into High Gear
It’s hard to keep track of all the news stories around poor AI decisions and fairness such as biased recruiting tools and discriminatory recommendations in healthcare. Reports have detailed how the pandemic nullifies existing AI to the point that leaders can’t trust predictions. Several high-profile cases of AI-backed services and regulatory compliance have led to large settlements like the $55 million settlement for mortgage discrimination. And most of us are tired of news stories about bias lurking in AI systems unbeknownst to the creators or users until something unacceptable happens.
The growing attention on making our AI systems perform better has fueled investment and activities in AI, especially in building Responsible AI. At Fiddler, we’ve seen accelerating customer deployments, market recognition, funding, and product innovation – all with the aim to build trust into AI.
There’s no doubt that responsible AI is shifting into high gear. But let’s take a breadth and go over key shifts that happened this year.
It’s clear that AI-based decisions are making their way into most businesses and most of our lives. A few years ago, 83% of companies believed AI was a strategic priority for their business. Yet, at the same time, 76% of CEOs were most concerned about the potential for bias and lack of transparency when it comes to AI adoption.
Despite this tension, the AI market continues to grow. In February, IDC projected that the global AI market will reach over half a trillion U.S. dollars by 2024. So it’s no surprise that companies are looking for ways to build their AI solutions more responsibly, and the interest in Fiddler’s practical framework has increased. In addition to working with a growing number of Fortune 500 companies, Fiddler raised an additional $32 million in funding earlier this year to accelerate product innovation.
In 2021 Fiddler was recognized by:
- CB Insights in their most promising AI 100
- Forbes in their AI 50 for the second year in a row
- Gartner as a sample vendor for Explainable AI in two 2021 Gartner Hype Cycle reports
- Hype Cycle for Data Science and Machine Learning
- Hype Cycle for Analytics and Business Intelligence
- Gartner as a representative vendor in the 2021 Gartner Market Guide for AI Trust, Risk, and Security Management
And just last month, the 4th annual XAI Summit provided a full day of community content on explainable AI and MLOps. We were thrilled to host about a thousand participants that joined to hear AI experts from companies that included Salesforce, Facebook, and AWS to DoorDash, JP Morgan, and U.S. Bank.
From the XAI Summit:
“We did a study of 2,400 consumers worldwide, 86% of which said they would be more loyal to ethical companies, 69% of whom said they would spend more with companies they regarded to be ethical, and 75%, three quarters of those consumers, would not buy from an unethical company.“
Yoav Schlesinger, Director, Ethical AI Practice, Salesforce
Fiddler Innovation in 2021
As customer demand and investment in AI startups increases (Dealroom predicts $90 billion this year, up from 60 billion in 2020), expect to see more use cases and breakthroughs in AI. For Fiddler, 2021 has been a whirlwind of product innovation, focusing on enabling users to operationalize ML and AI with trust and transparency.
First, we expanded our XAI functionality to include all-purpose explainable AI to explain various models, from simple tabular ML models to complex, multimodal deep learning. This XAI used our industry-first model analytics for global and cohort analysis in addition to speeding up troubleshooting.
Second, we brought the Fiddler Bias Detector to market this year to evaluate bias in static training data and dynamic live production data. And because fairness is often complicated, we’re proud to offer a way to assess intersectional bias where multiple attributes can overlap and amplify discrimination.
Third and most significantly in 2021, we brought together our capabilities in an enterprise platform for ML Model Performance Management (MPM). This integrated comprehensive ML model monitoring with built-in explainability and advanced bias detection in a framework with continuous feedback, centralized controls, and a unified dashboard.
The Fiddler MPM platform can ingest from any data source or model type with pluggable services. The flexible tech stack augments existing ML workflows from hooking into input/output logs for monitoring to consuming model artifacts and creating transformations for advanced explainability. In addition to supporting enterprise-scale, we added robust security and privacy features to meet stringent demands of Fortune 500 enterprises in financial services, healthcare, and other industries.
Bringing these capabilities together in a unified platform enables teams to conquer complex models and data pipelines while building trusted AI solutions with less bias. Centralized management and feedback for troubleshooting and facilitating continuous ML improvement.
Reaching More AI Practitioners
In our mission to build trust into AI, we need to reach more people; make the Fiddler MPM more accessible to more data scientists and MLOps teams. As part of this goal, we’ve made Fiddler available as a managed SaaS offering on AWS and on the AWS Marketplace as well. Customers can leverage Fiddler using a cloud-native experience and simplify procurement and billing with their existing AWS cloud subscriptions.
In addition, Amazon SageMaker is a popular choice for data scientists to build, train, and deploy ML models fast. As an Amazon SageMaker partner, Fiddler enables AWS users to accelerate existing projects with advanced monitoring, explainability, and bias detection. We’ve seen a fantastic amount of engagement with AWS customers as they put their ML models into production. This post on AWS explains how Fiddler Uses AWS to Make it Easy for Companies to Explain ML Models.
“Some of the strongest entrepreneurs today are people who have seen a problem in their previous role, begun to solve it, and now want to solve it for the world. That’s exactly what Krishna’s team at Fiddler has done.”Allie Miller, US Head of ML Business Development, Startups and Venture Capital at AWS.
Joint customers of Fiddler and AWS benefit from this collaboration that supports some of the most advanced AI companies with real-time business at scale.
“Bigabid provides state-of-the-art scientific advertising using AI at scale. Our rich data pipelines support thousands of machine learning models that we regularly optimize. Using Fiddler to maintain and improve our ML performance in conjunction with AWS services for reliable scale, means we can flex and react at the speed of business.”Amit Attias, Co-Founder and CTO at Bigabid
To meet the growing interest from the AWS data science community, we’re participating in various AWS events including the below:
- re:Invent: Stop by booth 1863 or check out our session on Tuesday at 2:40 PM titled, “Build ML Observability with Model Performance Management” in the Houdini Partner theater at the Venetian.
- AWS AI/ML Day for Startups: Continuous ML Improvement: Automated Monitoring with Built-In Explainability
- APN TV: Continuous ML Improvement: Why Observability Needs Explainability
With all this momentum, activity, and interest, responsible AI is definitely shifting into high gear. Are you ready? Stop by our booth at re:Invent to discuss your plans or check out the recordings from the XAI summit to get started.
And let me know how we can help!
Evangelist for Responsible AI