Eureka! on Steroids

October 2018: Christie’s auction house sells a work created by the French artist collective Obvious in collaboration with AI for nearly half a million dollars.

December 2023: A group of researchers reports in Nature that they have solved a previously unsolved mathematics problem by collaborating with generative AI (GenAI).

June 2024: Researchers at the University of Pennsylvania leveraged machine learning (ML) to discover nearly 1 million potential antibiotic compounds, with dozens showing activity against antibiotic-resistant bacteria.

October 2024: two Nobel Prizes were awarded for discoveries related to ML, both in physics and chemistry.

October 2024: Joëlle Barral, head of AI research at Google DeepMind, recently noted, “AI will accelerate research far beyond what we currently imagine.”

A series of AI ripples in the pond of creativity.

Will AI, indeed, augment R&D&I? Will the pace of scientific discoveries significantly increase thanks to AI? What role does AI play in the “Eureka!” moment? What role does AI play in the countless tasks that define R&D&I? These are some of the questions that Arthur D. Little’s (ADL’s) Blue Shift wanted to investigate.

5 key findings

1. There is no blanket model; however, AI augments researchers, across every step of innovation.

AI impacts both the productivity and the creativity dimensions, augmenting and empowering human researchers. Companies that quietly leverage AI, using general-purpose large language models (LLMs) and smaller specialized models, are already seeing 10x productivity gains in some situations. Choosing the right model for the right application is key, and sometimes AI is not the best answer: there is no blanket model. AI serves well as a coordinator between diverse digital tools like simulation, good old-fashioned AI (GOFAI), GenAI, graphs, rules & heuristics, and Bayesian networks, while keeping the human in the loop. This oversight role of AI will be increasingly key in a future where complex systems and "systems of systems" are the order of the day.

2. Successful integration of AI into R&D requires agility, careful prioritization, aligned organization, and, above all, robust data management.

Organizations need to be agile given the speed of AI development. They need to remain focused on solving specific high-impact problems, not just deploying AI. This means that strategic prioritization is key, identifying where the trade-offs between data availability, AI tool capabilities. and solution impact are most favorable. Make, buy, or fine-tune decisions are important — in fact, most core R&D&I problems lend themselves well to fine-tuning existing open source models. Organizational and governance models need to be able to access scarce data science talent and ensure alignment with IT departments to address security and compliance requirements while maintaining the necessary speed. Data is the game-changer, not algorithms. Data management excellence will be the differentiator as algorithms become increasingly commoditized.

3. R&D&I departments will increasingly rely on open source models, use-case specific wrappers developed by start-ups, and inference-as-a-services.

While the value chain for AI in R&D&I heavily relies on major open source models from players such as Meta, Microsoft, and Nvidia, smaller players, such as Mistral and Cohere, also form a key part of the ecosystem, as do academic institutions. Applications tailored for every part of the R&D&I process already exist, as do start-ups targeting vertical-specific problems, although many are not yet adopted at scale. The cost of implementing and maintaining sufficient computing power is large, but hosting providers are increasingly offering inference-as-a-service models, running inferences and queries in the cloud to remove the need for in-house infrastructure, lowering up-front expenses and democratizing access to AI.

4. Three critical uncertainties (performance, trust, and affordability) condition the future of AI in R&D&I. To strategize today, companies should be aware of an impact spectrum composed of six plausible future scenarios.

A breakdown of the three main critical uncertainties:

•  Performance — whether AI will meet the high bar necessary for many R&D&I problems
•  Trust — the extent to which researchers, developers, customers, and the public will trust and accept AI-generated outputs
•  Affordability — how far AI implementation will be constrained by costs, skills, resources, and environmental impacts

These lead to six plausible future scenarios on a spectrum between AI transforming every aspect of R&D&I at one end to being used only in selective, low risk use cases at the other. In between are scenarios reflecting different consequences for day-to-day R&D&I work, organizational evolution, and winners and losers. Recognizing these scenarios is important for R&D&I organizations as they chart a way forward for AI adoption, as shown in the infographic.
 

SIX SCENARIOS FOR 

THE FUTURE OF AI IN R&D&I

5. Organizations need to prepare now for the AI-based future by taking six "no regret" moves around compute power, data sharing, talent management, training, governance, and quality control

Whatever the future brings, organizations need to prepare for the transformative potential that AI offers R&D&I. That means taking six no-regret moves now. They need to mutualize compute power with partners to increase its affordability, encourage internal and external data sharing, better manage AI talent, train their workforces in AI fundamentals, reset data and AI governance approaches, and improve output controls over AI-generated content. 
 

Success will come from creating a balanced portfolio of AI-based R&D&I investments aligned with corporate objectives. This means considering the scope, costs and benefits of specific AI use cases and building the right balance of: 
 

• Hedging AI moves to ensure rapid response in case of future disruption
• Speculative high-risk/high-reward AI opportunities
• Shorting in future AI areas where AI impact is compromised by poor performance, low levels of trust, or excessive costs.