Many people feel that the world they know will never be the same because the realities around them are changing at an ever-increasing pace. Regarding the risk of war, pandemic, or other threatening events, Ulrich Beck’s description of the “Risk Society”1 we are living in might be accurate. But for business decision-makers, hiding by not informing themselves about these developments is not an option. They know that this can only be one side of the coin these days. Because when there are risks, there must be opportunities, which can also be represented in megatrends. Depending on the focus, they reach from sustainability over global geopolitical power shifts to urbanization2, and of course, technological disruptions3.
Of all these megatrends, artificial intelligence (AI) is currently generating the most hype. It is the most interesting one for decision-makers when they think about how their companies can use it in an efficient way. Even though modern AI is considered to be 10 years old4, the megatrend AI is still at the beginning, as worldwide investments in this technology are projected to reach 300 billion dollars in 2026 compared to 37.5 billion dollars in 20195. And it is the most backed sector by the VC fund industry6.
The current hype that started in 2022 is mainly because of OpenAI’s ChatGPT. It is the fastest-growing application of all time, based on the time needed to get one million users7. These numbers sent shockwaves through the business world, and giants like Google8, Alibaba9, and Meta10 are developing their own AI solutions. The current hype is mostly about generative AI, the ability of an algorithm to produce various types of content. That’s why you’re currently hearing all the stories of AI-generated video call transcript summaries, motivation letters, or even scientific studies. But how can you use AI technologies for your business?
Microsoft CEO Satya Nadella, for example, has the vision to develop an AI-driven copilot for everything. And they already have such solutions like background noise canceling in Teams or PowerPoint draft suggestions. Adept has a copilot for knowledge workers. Amazon manages its supply chains and instructs warehouse robots with AI.
The big tech companies found ways to sell some of those AI capabilities to customers of their cloud computing divisions, with rapidly growing revenue increases. There are robots sweeping up warehouses or AI helping sales teams to follow up on leads. As well as ML-trading bots. John Deere is looking into AI-generated “synthetic” data, which would help train other AI models, and in December 2021, Nike bought a firm that uses AI to generate new sneaker designs. Artificial intelligence will be applied to ever more jobs and company functions. Once trained, foundation AI models are good at performing a variety of tasks rather than a single specialized one11.
Nick Bostrom from Oxford University said that once something becomes useful and common enough, it is not labeled AI anymore. Databricks CEO Ali Ghodsi even called it “Boring AI”. It will integrate into and take over more and more jobs as well as company tasks. Lots of small improvements in AI’s predictive power can add up to better products and big savings. There are a lot of useful AI applications, and more importantly, they will only get more overwhelming in the future. They will be applicable in every industry, and their use case alignment and implementation into the running IT infrastructure will determine the future success of an organization. And in a business context, independent of the industry, AI can not only help improve process automation, but it is also about finding ways to implement this prediction technology in the best way possible. It is about the components of prediction and how human judgment fits into that equation12. And if this is managed properly, your company will have a bright AI-supported future.
A McKinsey Survey says that only 5% of companies are seeing large increases in earnings13. And according to a study by The European Business Review, only 11% of large European enterprises have the AI maturity needed to scale up quickly and effectively to reach their revenue goals14. This shows that it is challenging for a company to use AI efficiently.
During the previous AI hype in the mid-2010s, the technology did not take off exponentially as many business leaders expected. As for any technology, this has to do with the AI adoption curve. Namely, as good as it might be, it needs time to adapt to its environment and vice versa. So it is about managing your staff and technology, including your modern data stack, to get your business ready for the future. The problem is that even though the pressures to change are coming from all directions, it takes work for a company to decide where and how to start.
Or you might have started implementing some ERP or CRM systems, but because of minimal ROI generation, you have the feeling that you are on the wrong track and that there is something missing on your way to success.
The missing piece in all of this could be a data governance strategy. This would serve you as a base to manage better the data of your inventory and customer management efforts, and it would be a logical starting point to kick off your AI initiatives. And there is no better time to do this than right now.
As an umbrella term, data governance covers various functions of protecting and controlling data and data usage. We discuss leveraging the value of your data with data governance on our website and our data governance essentials whitepaper will help you tackle some of the issues of implementing AI efficiently within your data governance framework.
Having been trained on the internet, many foundation models reflect humanity with all the biases like discrimination – “When algorithms rule, values can wither”15. Ignite Meta, for example, pulled down Galactica, its foundation model for science, because it produced “pseudo-science”. This is because foundation models tend to be black boxes, offering no explanation of how they arrived at their results. They can create legal liabilities when things go wrong. This shows that the data quality issue is central to the artificial intelligence question because AI algorithms run on data. And they can be only so good as the quality of the data it is being fed allows it. So when implementing an AI tool, one needs to consider the ground truth, so the information that is known to be true and objective is what the AI has been trained and validated on16.
AI needs to be embedded in the right IT infrastructure, which provides it with enough computing power. It needs to be synchronized with a functioning (meta)data architecture at the data level. Besides this, new privacy and cyber regulations will affect AI and end organizational data hoarding. This will augment the importance of data-efficient AI techniques17, as well as bring up the need for data governance solutions that will not only help you comply with existing regulations but also get you ready to be prepared for regulations that will certainly come in the future. And it helps people in your company to think in a more structured data-cooperative way, allowing strategic data-sharing initiatives.
The possibilities that data governance can provide give you the benefits of restructuring your modern data stack and providing your staff with data literacy that goes well beyond AI. But it would be best if you implemented a functioning data governance strategy first before you can even think about AI. And Accurity offers great solutions for doing this. We not only assist our clients with a data quality and data observability solution that has helped banks comply with regulations, but Accurity also enables data mesh as it serves as an orchestration platform by providing adjacent tools such as a business glossary and data catalog and built-in connectors.
Let our product experts show you how the Accurity data intelligence platform can help you with your data governance and help prepare the way for your artificial intelligence initiative. Get your free demo now.
1 Beck, Ulrich (1992) Risk Society: Towards a New Modernity. Sage Publications. London
2 https://www.pwc.co.uk/issues/megatrends/
3 De Waal, Andre & Linthorst Julie (2022) Futurize! Dealing with Megatrends and Disruptions. A Handbook for the Future-Oriented CEO. Routledge. New York
4 c’t Magazin (2022) Happy Birthday, KI!: Künstliche Intelligenz: Vom Katzendetektor zum Maler, Texter und Gesprächspartner
5 https://www.idc.com/getdoc.jsp?containerId=prUS49670322
6 Forbes Middle East (2022) Startups: Top VC-Funded Industries
7 https://www.statista.com/chart/29174/time-to-one-million-users/
8 https://www.kdnuggets.com/2023/03/chatgpt-google-bard-comparison-technical-differences.html
10 https://medium.com/@bablulawrence/stanfords-alpaca-is-a-very-different-animal-c3c3abbd25fb
11 The Economist (2022) Information technology: The new age of AI
12 Agrawal Ajay, Gans Joshua & Goldfarb Avi (2022) Power and Prediction: The Disruptive Economics of Artificial Intelligence. Harvard Business Review Press. Harvard
14 The European Business Review (2022) Why mastering AI is harder than you think: And four essential behaviors to get it right
15 MIT Sloan Management Review (2023) When algorithms rule, values can wither: Building responsible AI systems starts with recognizing that technology solutions implicitly prioritize efficiency
16 MIT Sloan Management Review (2003) The No.1 question to ask when evaluating AI tools: Determining whether AI solution is worth implementing requires looking past performance reports and finding the ground truth on which the AI has been trained and validated
17 MIT Sloan Management Review (2023) How to build good AI solutions when data is scarce: Data-efficient AI techniques are emerging- and that means you don’t always need large volumes of labeled data to train AI systems based on neural networks