Please scan the QR code from your WeChat application:

Updates
These companies currently have access to the the Ethical AI Catalyst. Would you like the following company to keep you updated about their products and services by electronic means?
BolgiaTen Limited
BT Group plc
Vodafone Group
Description:

Please confirm that you would like to stop following this catalyst project.

 

Confirm Cancel

Log in

 

Not a user?

Register

0 Review(s)
0
Average based on 0 review(s)
5
0
4
0
3
0
2
0
1
0
URN
C22.0.375
Themes
AI , Data & Insights

What is the main problem you have identified:

Our fast and wide adoption of AI, AI innovation and democratization calls for proper safeguards to ensure that it's deployed responsibly and ethically. That is, 'trustworthy AI ', 'ethical AI by design' for enterprise. Who is affected? Every industry. For Telco operators, we have both extensive network and customer data, and various AI models and solutions, which is more urgent for proactively studying and managing this with vision and guidance.

How will this project team solve the challenge or problem:

We need a comprehensive study in order to: - Provide a set of sound and complete AI Ethics policies and guidance to ensure ‘Ethical AI By Design’ for enterprise - Golden rules to coach people mind-set to boost AI implementation and democratization (tell you a real story: recently the marketing guys from one of our LMs challenging us what’s the evidence why they should trust or believe the ML prediction results?) - (Optionally) potential research collaboration with world top AI institutes e.g. Oxford Univ. Cambridge, Imperial, etc. to keep tracking and update our Ethical AI framework and guidance with SOTA research fruits - The typical questions to tackle:

  1. How do we detect bias in data and/or AI model?
  2. How do we possibly detect and assess if/how much bias in vendor’s ‘black-box’ AI solutions?
  3. How to audit AI model, no more 'blackbox' ?
  4.  How to ensure Responsible AI?
  5. How to ensure transparent, explainable AI? 

 

How will the success of the solution be measured:

  1. A set of best practice and guidance to define and drive 'ethical AI by design' policies for enterprise;
  2. Enable trustworthy AI for production: explainable, fair and robust;
  3. AI becomes more transparent and explainable: AI model results can be explained and verified via test
  4. All AI models can be audit, no more 'blackbox';
  5. Historical/cultural/geo- biases in dataset can be detected and mitigated via effective means;
  6. Advocate and influence that AI cannot reach its full potential unless it can be accessed by everyone