Did you know the term “Artificial Intelligence” (AI) was first coined in 1956 at a summer conference by John McCarthy, an assistant professor of mathematics at Dartmouth College in Hanover, New Hampshire? It was the same year Elvis Presley released his first hit “Heartbreak Hotel” and the appearance of the first computer hard drive invented by IBM.

Fast forward 5 decades, so what’s in it for 2020?

According to Gartner 2020 CIO Agenda Survey, over 40% of leading organisations plan to deploy AI solutions by the end of 2020. They are expected to double the number of AI projects in place within the next year. On that account, Information Technology (IT) leaders across industries are enabling the evolution of AI pilots into scalable production and increasing business value realisation using AI pilots.

The potential to deliver real business value by leveraging AI initiatives is massive and kickstarting AI may seem easy. Many enterprises look for a mythical data scientist to write some codes and let the “magic” happen. However, this is not the case. Media outlets have portrayed AI as a magic “fix-it-all” solution, but AI is not something you can bolt onto existing systems and expect it to produce a transformational result.

In reality, many organisations struggle to determine a clear measure of tangible returns from AI adoption. The only way to realise tangible results and productivity from AI initiatives would be during the production of AI. However, most organisation find it difficult to scale AI pilots into enterprise-wide production due to skills, technology, and infrastructure challenges as depicted by Gartner in Figure 1. This limits the ability to realise AI’s potential business value impact on the organisation.

According to Chirag Dekate, Senior Director Analyst of Gartner — “IT leaders responsible for AI are discovering the “AI pilot paradox,” where launching pilots is deceptively easy but deploying them into production is notoriously challenging.”

Consequently, does this mean that you should stop investing in AI? Of course not.

According to PWC Global, the GDP of different countries are expected to grow with US$15.7 trillion potential contribution from AI by 2030, making it the biggest commercial opportunity in today’s rapid-changing economy. Moreover, Statista states that the global market of AI is expected to touch US$118.6 billion by 2025 and a study by Accenture shows that in 2035, AI-integrated countries have the ability to amplify their gross value added (GVA) by up to 40% compared to their baseline growth.

Technology in 2020 is moving at a breakneck speed. We could very well use AI to create new business value propositions to address current, new, and emerging issues that are strangling the industry or the society. For instance, using predictive analytics and big data to combat deadly disease outbreaks such as COVID-19 and SARS. The potential benefits are enormous.

 So, how then do we kickstart AI with success?

1. To have a bold digital vision in mind.

A vision is a vivid mental image of a desired future based on your goals and aspirations. It provides clear focus and broader perspective to think about or plan the future with imagination or wisdom.

Just like Amazon, its ultimate vision to create ‘The Everything Store”, a retail destination where people can buy everything they want, when they want, at the price they want. Instead of investing in an expensive marketing or advertising campaign, Amazon focused primarily on its customers and used predictions to determine future shopping trends. Thus, relying on positive reviews and word-of-mouth from customers, it made a mark in its industry. In just 20 years, Amazon has grown from an online startup selling books, to an enormous multi-platform, multi-industry technological disruptor, predicted to be worth US$3 trillion by 2026. Looking at Amazon’s fast progression and adaptation in this era of digital disruption, they might be able to see their digital vision fulfilled as the company constantly works boldly towards it.

In addition, did you know that Apple’s digital vision is to be the future of healthcare? In 1988, Apple made a video called “Healthcare 2008”.

This forgotten footage forecasts many of today’s innovations like telemedicine or electronic health records. Fast forward to today, Apple sees healthcare and wellness as a core part of its apps, services, and wearables strategies. Now the company is aiming to make healthcare more personal. It is evident with the release of the ability to download Apple Health medical records, and the Apple Watch ECG app with features similar to a single-lead ECG.

Leveraging on your bold digital vision, you are able to plan ahead with focus. This would help you determine a clear measure of tangible returns from AI adoption. Hence, bringing you one step closer to achieving your business goals. However, visions are simply fantasies unless backed up by action.

2. Closing the internal skills gap and have champions to set up a pilot project.

In order to kick-off the implementation of AI, organisations need to identify the internal skills gap related to AI and effectively address the gap to build experts to execute the AI vision. Acknowledging what you do and do not know about AI is crucial. Hence, bringing in external experts or AI consultants can be invaluable in piloting your AI project.

Bringing together a team of internal and external experts can be challenging. However, it will allow the team to have a tighter time frame and with good reason because it keeps them focused on straightforward goals. It is important to have both AI and business subject-matter experts to be part of the AI pilot project team in order to produce valuable results that may help the organisation to move the needle to a certain degree.

3. Cleanse, integrate, and store your data before AI adoption.

We have many data-rich but intelligence-poor organisations around us. Using out-of-date processes and legacy technology will limit one’s ability to make timely operational decisions. You are already generating the data that’s needed to produce valuable results. But without the right tools, you are not extracting any value from the volumes of data you stored. How then can we glean relevant, actionable insights and guide corporate strategy and decisions from these data?

By 2025, International Data Corporation (IDC) estimates that there will be 41.6 billion connected IoT devices, or “things,” generating 79.4 zettabytes (ZB) of data. Imagine if we could harness this tremendous amount of data and turn them into usable insights. However, the problem is that we just do not have the human capacity and hours to sort through so much information. Cleaning redundant and low-quality data, removing outliers that skew results, and interpolating missing records might not be appealing but buried within these assets are gems of information that can potentially provide us the key to unlock the future.

Therefore, to avoid the unnecessary risk of producing results based on inaccurate and unreliable information, it is crucial that organisations have clean data to feed the AI models. Moreover, you can produce high-quality results by effectively consolidating and organising your data sources into an optimised data lake or data warehouse for enterprise-wide data cleansing, curation, and storage for high-quality data.

4. Use a human-centred design approach to build perspective.

Human-centred research and design elevates artificial intelligence. It builds an inclusive AI world. Users are involved in the design process right from the very start. This enables users to create AI that brings value to solve some of the worst problems. In order to evaluate the true impact of its predictions, recommendations, and decisions, it is essential to assess the way actual users experience your system.

To produce an inclusive AI world, having clarity in AI ethics is crucial. It is important that the management anticipates compliance with upcoming laws and regulations, and possible impacts on individuals’ rights, society, and ethical values. This may very well help you avoid costly pitfalls and there is a profound humanity to it all.

Take the Cambridge Analytica data scandal as an example. It was a major political scandal in early 2018. Cambridge Analytica was a British political consulting firm which harvested the personal data of millions of people’s Facebook profiles without their consent and used it for political advertising purposes.

Image: The Facebook and Cambridge Analytica scandal in 2018.
Credit: Vox

Therefore, working out ethics in AI is not just a feel-good endeavour, transparency in the technology is important.

In addition to using a human-centred design, most organisations use AI for predictions and automation such as automating the decision-making process in awarding a loan. However, automated algorithmic decisions can reflect and amplify undesirable patterns with the data they are trained on. You see, AI systems feed off both positive and negative interactions with people, so we need to consider the work in a greater context of socio-technical systems. To ensure right implementation, one needs to model potential adverse feedback early in the design process, followed by specific live testing and iteration for a small fraction of traffic before full deployment.

It may cost you your business if it’s not being done right.

A vivid example is Tay, a chatbot released by Microsoft Corporation to Twitter, designed to learn about the world through conversations with its users. This innocent chatbot had to be switched off within 24 hours after pranksters trained it to post inflammatory and offensive tweets through its Twitter account.

Tweets by Tay, a chatbot released by Microsoft Corporation.

Lastly, build a rich variety of user perspectives and design a solution that focuses on your users’ pain points. AI exists to make human life simpler and richer. Thus, maintaining the ‘human element’ in the way it is made, delivered, used, and improved will most certainly make it a lot more useful and successful. To do so, you may incorporate feedback before and throughout project development. Fine-tune the product or your algorithms based on users’ feedback because even the best AI will become quickly redundant without inputs from real humans on how to accelerate strategic decisions and processes.

In a nutshell, to kickstart AI, one must have a bold digital vision. Having said that, embracing digital transformation comes before jumping into the AI bandwagon. Organisations should look at AI through the lens of business capabilities rather than technologies. This games out possible consequences of each action and streamline decision-making processes. So, if you are yet to embrace digitalisation in your company, what are you waiting for?

JURISTECH: Digital enablers that work behind the scenes tirelessly

References:

1. https://www.gartner.com/smarterwithgartner/gartner-predicts-the-future-of-ai-technologies/
2. https://www.gartner.com/smarterwithgartner/the-gartner-2020-cio-agenda-winning-in-the-turns/
3. https://www.europeanbusinessreview.com/building-the-everything-store-amazons-cycles-of-creativity-and-circles-of-destruction/
4. https://www.pwc.com/gx/en/issues/data-and-analytics/publications/artificial-intelligence-study.html
5. https://www.statista.com/statistics/607716/worldwide-artificial-intelligence-market-revenues/
6. https://www.accenture.com/my-en/insight-artificial-intelligence-future-growth
7. https://www.advisory.com/daily-briefing/2019/01/22/apple
8. https://www.yourteaminindia.com/blog/artificial-intelligence-accelerate-business/
9. https://www.theverge.com/2016/3/24/11297050/tay-microsoft-chatbot-racist
10. https://ai.google/responsibilities/responsible-ai-practices/
11. https://xo.xello.com.au/blog/6-essential-steps-to-prepare-for-ai-adoption
12. https://www.gartner.com/document/3976201?ref=solrAll&refval=235826587
13. https://www.gartner.com/analyst/60395?qid=caf83e9f7dfc4e4ed5f800deb5