Augmented AI – Impact on society as a result of the AI ethical index

We are living in a society surrounded by gadgets and technologies. Humans and devices are constantly generating tremendous amount of data footprints. As such, technology is advanced so that we can augment data footprints to teach machine (ML) to help humans’ takeover repetitive tasks. Key question in today’s world – are we using ethical set of data footprints before we augment into machines?

Automation and augmented AI are bringing new opportunities and greater efficiencies to both businesses and society. Few such examples are extracting camera (CCTV), device, sensor, information available on public domains (google search) and social media etc. Feeding information from such channels (data sources) is easy to plug-in but it has bigger implications on society if it’s not done in controlled and ethical manner and through the lenses of data governance. Do we really ensure that information we pull and process to convert into augmented AI is reliable and safe to consume before integrating into massive application?

Do we augment ethics to AI programs as a result of good data governance?

Let’s try to understand from simplified visualisation to compare visualisation accuracy between human vs machine. Following example showcases complete cycle starting from evolvement, intelligence development and path to maturity. This figure illustrates on how artificial intelligence (AI) matures over a period using iterative process and historical data. Eventually intelligence plays a key role in defining direction to have positive or negative impact on society.

Conclusion:

1. Data plays a vital role while developing intelligence in machines, hence augmentation of ethical data is important
2. Data governance is key part of software development and ethics are mainly driven by organisation policy. e.g. GDPR, and Personal Data privacy laws

How to measure such data governance and data ethics compliance?

1. Data Analytics plays key role to give insights to evaluate augmented ethics, and to do predictions of directional growth. Monitoring of ethical index is important factor while augmenting AI to application at scale
2. Good data governance brings positive impact on society, vs poor governance brings disaster at occasions

In summary, ethics and governance have always been one of the hot subjects while deploying AI, as it has direct impact on humans and society. At the same time Data Analytics plays a vital role to compare “desire to build vs actual act of the machine”.

Argument and opinion

If this is complicated and results are unknown for society then why we augment AI at all?

There are better examples where machines have made our life easier than we thought through Augmented Intelligence (AI). Example exercise to understand about human interprets information vs capability of machine interpretation using machine learning (ML).

As we know in today’s world data is growing more than ever including unstructured data for an example. Every information creates unique digital footprint in destination database. While processing such raw data, developers use APIs available from the marketplace or Github open source repository. Most software APIs convert raw data or forms modernised form of structured data and store into data lake. Then it is consumed through augmented AI by different organisations for various purposes. Hence why it is important to apply ethical and governed intelligence which creates a positive footprint on stored data rather than destructive set of actions.

Example of such intelligent APIs to showcase how it extract information from unstructured data and shows augmented AI confidence level in percentage.

Over the period I have done various deployments where data reliability (data integrity) from original source has been a key challenge for the organisations. Scale is very important parameter while deriving social impact. Hence deriving ethics parameter and social impact for augmented AI is comparatively easy for private businesses as such data is tapped from known sources (within controlled zone), whereas for the large-scale government organisations it’s difficult to assess implication on society and ethics index. Most recent examples are Global Data Privacy Regulation (GDPR), Crypto regulations etc.

There is an example for self-assessment purpose to understand why ethical approach in AI augmentation is required and relevance of data integrity:

1. There are many secure applications we use in our mobile device, handles our private and transactional data in secure and most ethical manner. We get product recommendation based on our taste, shopping pattern, search history etc. This example demonstrates ethical AI augmentation
2. What if we use AI to read social media news, articles, google search and live stream media? What happens if we generate impression purely based on key words or images? Can we consider this as ethical and governed data for AI augmentation?

Someone said it well – “Lies can become truth, if we let them”. In this information edge how we can verify the authenticity of every information passed through? Someone may have an opinion that’s pragmatically difficult ask, but can we resolve this challenge with help of high index ethical AI engine to help human to identify authenticity of data before augment at a scale? Extensively, AI can help identify true source of information and ethical data sources from the internet to benefit society.

At Capgemini, we constantly strive to solve such puzzles and exploit technology every other day. We are encouraged and supported to participate in global initiatives to work collaboratively with our colleagues across the globe. One of the recent initiatives where we exploited technology to do example experiment is Insights and Data AI Academy through the AI Guru program.

Kevin Ramani

Data Scientist and Product Development

Kevin has over 15 years of technology experience working with Product development team, Service delivery teams and Center of Excellence (CoE). Kevin has extensive experience enabling innovations, accelerator programs and product development for large scale customers.
Outside work Kevin has hobby to write blog, develop ideas and exploit technology.