The 2023 Digital Decade Report was published on September 27th, 2023 by the European Commission and describes in detail the state of the European Union in its progress to digital transformation. The goal for Artificial Intelligence adoption by Enterprises is only at 11% which confirms the skeptic opinion that, even with the latest hype around LLMs and GenAI, companies are still unsuccessful (or unwilling) to adopt AI at scale.

The report (published here) evaluates the European Union’s (EU) advancements in its digital transformation journey, in line with the benchmarks set by the Digital Decade Policy Programme 2030. The report delves into the recent shifts in digital policies and highlights the EU’s progress in meeting the set goals. It provides a clear picture of where the EU currently positions itself at the commencement of the Digital Decade Policy Programme’s execution phase.

The target set by the Digital Decade for Europe for AI take-up by enterprises is 75%, but the report shows that with the current trends, by 2030 only 20% of Enterprises will have adopted Artificial Intelligence.

Contrary to the recent “hype” escalated by the emergence of public Large Language Models (LLMs) such as ChatGPT claiming that revolutionary digital transformation with AI is imminent, the report shows that progress is indeed very slow. EU member states are encouraged in the report to join forces under the European Digital Infrastructure Consortia (EDIC) or other schemes to jointly build cutting-edge Europe-based AI models, possibly also through the proposed EDIC in the field of language technologies. Looking at Venture Capitalism, the best EU ecosystem – Berlin – was ranked 13th
worldwide followed by Amsterdam (14th), and Paris (18th). The situation is even more critical in deep tech, including AI, where EU venture capital is still far behind the US. Another important aspect of the report is that there are 10 EU countries with AI adoption below 5%, which is a third of the total number of EU countries.

What are the causes of low AI adoption in European enterprises?

The main cause can be correlated with the level of digitalization overall, but Artificial Intelligence is now the only technology that has non-stop undiscerned media coverage fueled by the rapid progress of LLMs and Generative AI. Without looking into the facts, one might think that companies are taking up AI with a speed never seen before, processes are automated every day, and we are witnessing an “AI Revolution” comparable to the 3rd Industrial Revolution. The reality is that – beyond the hype – AI can bring immense and irreversible progress, but adoption is not something that happens overnight, for a number of reasons described accurately by research group AI Multiple in the article Enterprise Generative AI: 10+ Use cases & LLM Best Practices. In summary, Generative AI, AI, or any other technology adopted by an enterprise would need to adhere to the below checklist:

✅ Consistent
Enterprise customers need predictability and enterprises deliver that. This sets them apart from immature businesses.

✅ Controlled
Building with evolving 3rd party APIs is building on sand. Enterprises need to own at least parts of the tech stack.

✅ Explainable
Enterprise users need to know the data that drive decisions. RAG can support this.

✅ Reliable
Through human-in-the-loop or guardrails, expensive mistakes need to be avoided.

✅ Secure
Depending on the attack surface, securing a model can be a trivial or complex but it needs to be considered.

✅ Ethically trained
An LLM built on unethical data is a bomb waiting to explode. Enterprises need to understand the training data.

✅ Fair
Bias in training data can impact model effectiveness.

✅ Licensed
LLM licensing is complex but important. You don’t want to rely on Llama-2 in a product that will have 700M active users next year.

✅ Sustainable
Business leaders should be aware of the full cost of generative AI and identify ways to minimize its ecological and financial costs.

https://research.aimultiple.com/enterprise-generative-ai/

The same approach is recommended by AI veteran Andrew Ng:

Looking for AI project ideas for a large company or industry sector I share in The Batch a brainstorming recipe: (i) break jobs down into tasks, (ii) evaluate common tasks for AI assistance or automation, (iii) assess value.

https://www.linkedin.com/feed/update/urn:li:activity:7112147813479677952/

These common-sense approaches to adopting AI should also be driven by a Consulting approach, where expectations are correctly set, goals are accurately defined and the customers are fully engaged in the process.

Text Analytics is now the most known AI discipline, thanks to ChatGPT’s amazing capabilities. But this also sets unrealistic expectations for real-world projects, where public LLMs’ capabilities to answer questions based on Wikipedia and public web content are not sufficient. Regulated industries, industrial, chemical, and automotive domains have very specific needs, and uncovering the best use cases for AI, and the uses with the most business impact, is a task in in itself. This can only be accomplished through an AI consultative approach, where business analysis, machine learning skills, and data science are combined together to understand the customer’s processes and market position, in order to identify realistic AI use cases that will bring value. As opposed to ad-hoc adoption of public AI systems, fragile prompt-based integrations have serious data privacy issues.

The Digital Decade report is a realistic view of where Europe is, and where it needs to be, but offers little on what are the steps that need to be taken to move forward. Luckily, the Industry is moving forward despite the adversities, the competition from the US and China, and new successful AI implementations are being delivered consistently by European companies that go beyond the hype, and actually bring value to the AI ecosystem.