Big Ideas on productivity and personalization for Data and AI in 2023

The impact of bringing together data and artificial intelligence (AI) will be at the center of 2023.? The potential for this combination of technology to accelerate innovation and create competitive advantages is something companies can use to get ahead or they'll get left behind.?

We’ve pulled together a few big ideas that we think can help leaders understand the role these technologies will increasingly play in driving business growth and innovation through the uncertain economy.?

Generative AI will find its voice, and it could help you find yours

Generative AI can streamline many time intensive tasks for people and companies. In practice, the technology can show up in various ways – from creative applications of text-to-image translation to participating in helpful dialogue within a chat feature or automating complex tasks with basic instruction. The technology is evolving quickly, becoming less abstract and more accessible every day. Gartner included generative AI in its Emerging Technologies and Trends Impact Radar for 2022 report as one of the most impactful and rapidly evolving technologies that could bring about a productivity revolution. Large AI models, like DALL-E 2, GPT-3, ChatGPT, Stable Diffusion, Gato, Lamda and many more - have shown enormous potential, leaving us to ask “what’s next?”?

In 2023, we’ll see companies pushing the boundaries of what’s possible with generative AI to accelerate productivity for their teams and users. For example, companies might start using this technology to help their marketing teams create the right copy and creative, a challenging task in a world of near infinite audience tastes, or a teacher could use it to create more engaging content to captivate their class. This is just scratching the surface of what’s possible and we’re excited to see how generative AI can unlock business productivity in entirely new ways in 2023 and beyond. At the same time, we will have to rethink how we evaluate skills: teachers will have to rethink essay writing assignments, and companies will have to rethink evaluating coding skills of engineers.?

Large-scale modeling algorithms and infrastructure will become more powerful together

Large-scale models are incredibly powerful, but the algorithms that leverage those models are nothing without the right infrastructure. As companies make bigger leaps in what it means to be “large-scale,” infrastructure investments will need to be well thought out and able to complement the modeling work for the long-term.?

The urgency in investment is proving very real for organizations. In McKinsey’s The state of AI in 2022 survey, a key finding from the survey was that the level of investment in AI has increased alongside its rising adoption. And when looking ahead, McKinsey found that 63 percent of respondents said they expect their organizations’ investment to increase over the next three years.

In 2023, companies will need to invest in ways to tie infrastructure and algorithms for large-scale modeling closer together. Since many companies made decisions on infrastructure before considering their large-scale modeling needs, data and AI practitioners will need to decide where they can optimize their current infrastructure and when it’s time to make dedicated investments in new infrastructure. We’ll start to see hardware dedicated to large scale modeling opportunities gain a foothold in companies’ infrastructure roadmap.

AI models on the fly will accelerate personalization

There has often been a tradeoff between quality and speed, especially when exploring an existing problem or testing something new. With data science and AI, this can mean big changes in the intended outcomes.

Next year, we will see big steps towards making AI models adapt on the fly. With the right approaches to platform models, embeddings, proper infrastructure, and data fidelity, more AI models will be able to generate outcomes and be retrained at real-time speeds. This will be felt in clear ways with better personalization and a meaningful accounting of real-time context. Businesses and people will get closer to having AI deliver the right content to the right users at the right time and place.

Of course, the benefits of AI, like personalization and real-time adaptation by AI models, depend on data. Companies that plan to build AI solutions must take steps to ensure they have the right permissions in place to use the data for its intended purpose and to prevent biased outcomes as a result of the data used.

Companies will lean into taking responsibility for the ethical and responsible use of AI

AI is taking center stage, as companies are considering new ways to use it to scale their businesses and offer value to their users and customers. Because of AI’s increasing pervasiveness, governments and organizations around the world have started looking into policies to help shape the ethical, trustworthy, and responsible use of the technology, for example the proposal for an AI Act in the EU or the Algorithmic Accountability Act of 2022 in the U.S. Congress. But that process takes time and, in the interim, many companies are taking ownership of applying AI ethically and responsibly, and sharing principles, best practices, tools and case studies. Examples include Microsoft’s Responsible AI Standard and Google’s Responsible AI Practices. Some companies, like Anthropic, have recently been founded as public benefit corporations with a mission to build “reliable, interpretable, and steerable AI systems” and with focus on large language models and generative AI.

For many companies, responsible AI will include adherence to core principles such as transparency, fairness, and accountability. It also can include supporting a culture that empowers AI practitioners to question, test, and evaluate the ethical use of their AI models with explainability and bias testing in mind.? Taking a proactive step in managing those aspects of AI use will help prepare for changes in future regulation while still delivering business benefits and innovation. Ultimately, building and deploying AI responsibly is not a zero-sum game: it helps build better products and services for everyone.

2023 will be an incredibly exciting year for both data and AI. We expect to also see organizations support new standards, while seeking opportunities to adapt, innovate and improve. It’s an ambitious task, but we are optimistic and curious to see where both spaces head into 2023.?

What are your Big Ideas for the year ahead? Let us know in the comments! #BigIdeas2023