For Nicolas Griedlich, AI & Data Leader at Deloitte Luxembourg, the new technology has the ability to disrupt industries and generate a competitive advantage for the companies that harness its power. However, responsible implementation and governance will be critical.
On 7 December 2023, Deloitte Luxembourg hosted the 5th edition of its Disrupted/Disrupter Conference dedicated to Generative AI, at the Luxembourg House of Financial Technology (LHoFT).
Nicolas Griedlich (Partner | Artificial Intelligence & Data, Global ESG FSI Tech leader at Deloitte Luxembourg) explains why and how Generative AI can bolster various company functions in Luxembourg and how it can significantly contribute to the country’s GDP growth.
Nicolas Griedlich, in today’s times of fake news and content, disinformation at a large scale, how trustful are the data used to generate content?
In recent news, the public has been exposed to a series of deepfakes, including a bizarre image of the Pope donning a puffer coat and a premature depiction of Donald Trump’s arrest. These instances underscore the importance of traceability in content production.
As we navigate this digital age, the need to critically evaluate information, particularly from unfamiliar sources, has emerged as a best practice. Consumers are increasingly demanding data trustworthiness in content generation, which raises questions about the reliability of sources used to train Language Model Machines (LLMs). Ultimately, the crux of any data analysis lies in ensuring the use of appropriate and authorized data sources.
How ready are Luxembourg’s companies to explore and implement the use of Generative AI in their business?
First and foremost, businesses across various industries heavily rely on documents, both paper-based and digital, as integral parts of their operations. These documents are often analyzed or used to extract pertinent information for a range of purposes, from exploratory research to decision-making support and risk oversight.
With Generative AI’s remarkable ability to process and explore such content, Luxembourg-based companies can leverage this innovative AI model for their unique use cases.
Technologically, Luxembourg is well-prepared for the advent of Generative AI. The nation boasts of advanced infrastructure like the Meluxina supercomputer, which can process vast amounts of data within the Luxembourg network. Additionally, cloud hyperscalers offer scalable storage and computing capabilities to support Generative AI use cases. Companies like Nvidia even provide the flexibility to start on-premises and transition to the cloud when ready.
However, challenges persist, particularly in terms of adoption. Generative AI needs to be not only explainable but also comprehensible to end users who will incorporate it into their daily tasks. The effectiveness of chatGPT-like systems, for instance, depends on the quality of user prompts.
While some individuals excel at articulating their needs, others construct their thoughts iteratively, requiring collaborative effort to reach meaningful outcomes which can lead to irrelevant or partial answers from a Generative AI-enabled chatbot.
Another potential hurdle is the ambiguity surrounding the terms of use for available LLM models, which could hinder Generative AI adoption. Like any software, Generative AI packages must clarify how their LLMs are built to ensure legal compliance and how they will handle client data.
The forthcoming AI Act regulation is expected to address this issue by mandating AI model assessments to prevent unauthorized activities.
In which industries (banking, insurance, industry 4.0…) more particularly, and in which company fields?
At present, primarily Financial Services Institutions (FSI) and European institutions are exploring the use of Generative AI for the aforementioned use cases. This technology can bolster various company functions, be it operations, finance, risk management, or support functions. Any process that involves manual document handling can potentially benefit from Generative AI.
However, the key lies in ensuring a positive business case. The primary considerations include compliance with internal guidelines and the AI Act, data availability, and the ability to deliver expected outcomes that expedite the end-user journey.
Through these services, clients are empowered to leverage Generative AI, driving business process improvement, informed decision-making, and ultimately, fostering growth and innovation.”
Where, how, and by how much do you see the country can increase its GDP thanks to Generative AI?
As said before, Luxembourg is well-positioned to embrace the disruption brought about by Generative AI. The country boasts a robust technological infrastructure, including the Meluxina supercomputer, and a thriving ecosystem of cloud hyperscalers that provide scalable storage and computing capabilities.
This infrastructure, coupled with a strong commitment to digital innovation, makes Luxembourg ready for Generative AI.
We can expect that Generative AI could significantly contribute to Luxembourg’s GDP growth. By automating and optimizing various business processes, Generative AI can increase productivity, reduce costs, and drive innovation, all of which can stimulate economic growth.
How concretely does Deloitte Luxembourg help its clients? For what challenges and what needs?
Deloitte offers a comprehensive suite of services that span the entire implementation process. The journey begins with pinpointing potential opportunities. Experts work closely with clients to identify areas within their operations – be it finance, risk management, or support functions – where Generative AI can bring about transformative change.
Data, the lifeblood of any AI model, is the next focus. Clients are assisted in assessing the availability and quality of their data, a critical step in training effective Generative AI models.
Compliance is key in this digital age. The team ensures that the proposed use of Generative AI aligns not only with a company’s internal guidelines but also with industry regulations and laws such as the AI Act.
The creation and training of the Generative AI models are handled by a team of AI experts, who fine-tune the models for optimal performance in their specific use cases. But the journey doesn’t end with the development of the models.
Seamless integration into the client’s existing systems and processes is also facilitated. To ensure the models are used effectively, training is provided to end-users, equipping them with the knowledge to harness the full potential of the Generative AI models.
Finally, ongoing support and maintenance are offered, ensuring that the Generative AI models continue to operate effectively and deliver the expected outcomes.
