At the recent AI conference – NeurIPS – the focus was strongly on the emerging field of AI Ethics (in funding, development and application) and attempts to make AI more explainable and more transparent. Both ethics and explainability are two areas of weakness of the current “connectionist” AI system which rely on massive computation over big data to arrive at statistical patterns; patterns which are not accessible (or explainable) to their creators or their customers.
This lack of transparency is very troublesome for safety critical (driving), financially critical (credit card fraud detection) or socially critical (immigation or crime detection) applications. AI not being able to answer the simple question “why was my credit card blocked?” In a meaningful way is a big problem for building consumer trust in AI systems. GoogleCloud is a key player in cloud-based AI processing and is the backend of many commercial AI applications. Recognising the increasing negative press for unexplainable AI decisions that have societal impact, Google has launched a new service as part of its AI Suite of tools Explainable AI (https://cloud.google.com/explainable-ai). The toolset offers AI developers tools to apply before and after deployment of their AI-based applications.
Google is also under fire for influencing the path of core AI research due to its close relationship with many university research programmes and academics allegedly influencing research projects to report positively about AI impacts and development practices. NeurIPS speakers likened this overreach to the role of big tobacco companies in influencing public discourse on the health impact of smoking in the the US in the 1970s (https://venturebeat.com/2020/12/11/what-big-tech-and-big-tobacco-research-funding-have-in-common/). The AI industry is surely going to be facing challenging voices in 2021 as its aggressive application impacts more and more people’s lives.
Global Digital MOJO specialises in using AI to support digital transformation for our clients. AI for us is not only a buzzword; we develop AI and machine learning for use in our user behaviour tracking, games, education services, simulations, business matching platform (@connect) and in solving potential challenges for society. We use a design-led approach to balance design made by humans, and predictions made by AI analysis.
Please check out our related cases here if you are interested in digital services and solutions using MOJO AI.
MOJO Academy about AI application
MOJO cooperate with HKDA for AI business matching website building
Welcome to contact us if you need services support for AI business.
