2024 Trends: AI

AI, AI, AI. Nowadays, it’s impossible to go a day without seeing generative AI in headlines or hearing about it in podcasts, meetings, or conversations. For CEOs and business leaders, the abundance of online information about gen AI (mostly opinions and subjective projections) has obscured the present-day reality of the technology.

As we head into a new year, instead of continuing to guess or panic, leaders should seek to understand the real use cases, best-practices and risks so they’re well-positioned to take advantage of the many AI applications in their operations both now, and into the future.

To understand the current state of AI, it is essential to trace what’s happened over the last year. Let’s go back to November 2022, when ChatGPT was released, causing a sensation among business leaders outside the high-tech software industry. Since then, the proliferation of AI has been chaotic, with media, startups, venture capital firms, and tech leaders all voicing their opinions on its societal impact. The discussions have ranged from the potential end of humanity to AI being the solution for complex problems like high-energy fusion.

Despite the debates, AI has become a ubiquitous term in everyday life. This is evident from the staggering numbers, such as the 1.5 billion monthly visitors to ChatGPT, over 100 million daily active users on Bing AI, and the growing user base of Google Bard. These figures reflect a significant increase in the usage of AI models compared to their early versions just a few years ago.

Our AI & ML experts at Responsum and iOLAP (now Elixirr Digital) want to assist business leaders in transitioning to strategic generative AI action in 2024. This means avoiding unnecessary and random spending. Continue reading to gain insights from Partner Steve Steinberg and Principal Adam Hofmann on the main trends they observe in AI for the upcoming year and to learn about the key areas you should consider to harness this technology’s powerful and expanding capabilities effectively.

Master current best practices

We expect that the interest in artificial intelligence, specifically generative AI, will continue throughout 2024 as industry leaders move prototypes into production and corporate best practices emerge. To ensure that your business remains agile and adaptable in the innovative field of artificial intelligence, consider the following:

• Strategic data management: Large language models (LLMs) perform best with high-quality, well-organised data. While each use case may have specific requirements, it is important to make the data easily understandable for the model. Even if you are unsure about the best use case for generative AI in your business, improving data governance and cleanliness is always beneficial (PS: this is another excellent use case for AI).
• Identification and alignment: How is your business approaching the application of AI? Who is responsible for overseeing this process? Are all applications adequately aligned with strategic priorities? Is the sourcing of ideas done through a top-down or bottom-up structure? We advise establishing a solid, repeatable framework for identifying AI opportunities.
• Prioritisation and scoping: How are different AI applications prioritised? What types of AI are suitable (e.g., knowing when to use an LLM or a predictive ML model or when AI is unnecessary)? Who makes decisions about prioritisation? Does the list of criteria prioritise ethics, compliance, privacy, and security, or is it primarily focused on feasibility or business performance? Having well-defined, weighted criteria allows business leaders to implement AI technologies systematically over time.
• Start with PoCs: Many organisations are eager to start with a big, impressive use case. However, this approach will likely result in a significant budget expenditure without achieving a viable solution. Instead, it is recommended to prioritise investments in small, testable use cases. These can help you learn and gradually develop disruptive end-to-end solutions.
• Measurement: Like any business investment, measuring outcomes is crucial for shareholders and management. While the full impact of AI can be challenging to quantify, prioritising outcome measurement ensures tangible results and provides opportunities for iteration based on comparable data.
• Training and upskilling: AI is highly disruptive and often requires new skills. Workforces will need to be upskilled to handle new processes, and considering this early on is essential for smooth transitions and ensuring that all employees feel confident about their future with the company. Adequate preparation of the right people, processes, technologies, and information is necessary to minimise future delays in operations or workflows.

Incorporate AI into your organisational structure

It’s natural to feel uncertain about the practical next steps of implementing AI across your business, especially when it comes to responsibility and determining who will oversee these actions. Many companies address this organisational challenge by establishing AI Centres of Excellence (CoE).

An AI CoE has several responsibilities and should be at the core of any company, guiding all business functions such as IT, sales, finance, customer service, legal, and others. According to a 2019 survey, 37% of U.S. executives from large firms had already established similar internal organisations, which is expected to increase significantly in 2024.

By 2024, AI CoEs and new executive positions like Chief AI Officers are expected to become commonplace and play critical roles in transforming companies into more intelligent organisations.

Monitor emerging AI-related legislation & regulation

AI-related legislation is being developed and rapidly emerging worldwide, with the EU, US, and China taking the lead. Various organisations have released AI legislation trackers to inform companies about regulatory advancements. Examples include the Stanford University Human Centred Artificial Intelligence (HAI) and the International Association of Privacy Professionals (IAPP) trackers.

Companies that implement AI models in any capacity should be aware of the proposed and pending legislation to avoid the need for significant changes or overhauls to their models in the future. The main topics covered in these pieces of legislation revolve around ethics, security, and privacy. Companies need to review the abundant online resources available for each category before the year starts to ensure compliance with future AI-related actions.

Security & privacy lies at the core of AI

AI plays a dual role in the field of cybersecurity and data privacy. On one hand, it can enhance corporate capabilities in anomaly recognition, fraud and phishing detection, and efficient threat identification on a large scale. However, threat actors can also exploit AI to evade traditional security measures and expand their attack surface. As open-source AI becomes more accessible and powerful AI models gain familiarity, comprehensive cyber hygiene measures are essential in 2024.

The impact of AI on cybersecurity extends beyond business performance and stability risks; it also poses regulatory risks. Both the US and the EU have legislation in place that imposes specific cybersecurity screening requirements on foreign investments. These measures aim to safeguard critical economic infrastructure, sensitive data, intellectual property, and government operations. In 2024, conducting a thorough corporate cybersecurity audit is crucial to address the growing threats posed by AI and ensure compliance with regulatory standards. For more on cyber trends in 2024, click here.

Futuristic AI applications are now becoming a reality

Many companies view AI as a tool for improving internal efficiency, such as automation or faster document retrieval. However, AI technology has impressive applications in complex corporate divisions, including R&D. Sophisticated neural networks, a subset of AI, can tackle data-heavy and interconnected problems. They have already successfully solved challenging tasks like protein folding (as seen in Google DeepMind’s AlphaFold). In 2024, companies should broaden their understanding of how AI can be applied to their business. They need to recognise that even larger and more complex problems can now be solved through AI.

Prepare to adapt and evolve as the pace of change accelerates

AI technology is expected to advance further in 2024, driven by private investment and increased academic focus in universities. One notable shift in the evolution of AI is the transition from predictive to generative capabilities. An example of this shift is the growing popularity of LLM’s in 2023. As the transformer architecture and compute capabilities continue to improve, this trend is expected to persist and accelerate.

Furthermore, AI capabilities such as language, vision, and sound are converging into integrated models known as multimodal models, such as GPT-4V and Google Gemini. These systems offer a wide range of opportunities for human-machine interfacing and will undoubtedly change how we work even further. Organisations can explore automation and enhance internal efficiency by leveraging multimodal capabilities and customisable models like the new GPTs.

Despite the many AI considerations we’ve explored, our main recommendation is to start the learning process by experimenting with various generative AI use cases in your business. It is not necessary to build everything at once. Our experts can help you get started or navigate the complexities of implementing larger, more disruptive projects.