TL;DR:
- Dell’s CTO, John Roese, emphasizes the importance of data management in adopting AI models, highlighting the need to curate data to avoid biases.
- Unstructured data is key for large language models (LLMs) as they create connections independently, making it crucial to ensure data is strategically positioned for training.
- Most firms are currently unaware of the data requirements for AI, which is a concerning issue.
- Generative AI should be seen as a learning experience to prepare for the disruptive force of quantum computing.
- Quantum machine learning holds immense potential and is the logical progression beyond generative AI.
- Maintaining cybersecurity in the face of quantum computing advancements will be crucial as encryption standards may become vulnerable.
- Businesses must use the challenges of generative AI as a dress rehearsal for the quantum computing revolution.
Main AI News:
Businesses around the world are increasingly recognizing the immense potential of artificial intelligence (AI) and its role in shaping the future. However, as the global chief technology officer (CTO) at Dell, John Roese, points out, we must not lose sight of the looming threat of quantum computing. In an exclusive interview with ITPro at Dell Technologies World 2023, Roese sheds light on the challenges businesses face when adopting AI models and emphasizes the need to prepare for the quantum computing storm.
While many perceive the computing bottleneck as the main obstacle hindering the widespread deployment of AI, Roese believes that data management is the critical issue enterprises must address. To train large language models (LLMs) effectively, businesses must ensure access to the right data and establish robust data management infrastructure. Unfortunately, most organizations have not dedicated enough effort to this crucial aspect.
Roese cites Dell’s own exemplary approach to data management, highlighting the company’s four-year-long endeavor to eliminate non-inclusive language from its content library and internal code environment. By removing biased labels such as ‘whitelist’ and ‘blacklist,’ Dell actively mitigates the risk of unintentionally incorporating biases into its AI models. Roese warns that failing to curate data can lead to AI services that reflect inherent racism or misogyny, as exemplified by Microsoft’s infamous Tay scandal. He emphasizes the importance of creating datasets that align with an organization’s values to avoid such pitfalls.
In the realm of LLMs, unstructured data plays a pivotal role. Neural networks thrive on creating their own connections, disregarding arbitrary structures. Roese advises businesses to ensure their data is strategically positioned to facilitate effective training. By ensuring data is in the right place from the outset, organizations can save substantial time and resources that would otherwise be spent on restructuring data down the line.
Unfortunately, the majority of firms are currently unaware of these essential requirements. Roese expresses concern over this lack of awareness, stating, “No, they’re not, and that’s very disturbing, to be perfectly honest.” It is imperative for businesses to educate themselves about the necessary data management practices to unlock the true potential of AI.
While AI continues to captivate the world with its transformative capabilities, Roese urges organizations not to lose sight of what lies beyond generative AI. He suggests viewing the current upheaval caused by generative AI as a valuable “learning experience” to prepare for the disruptive technologies of the future. Roese emphasizes that generative AI is merely a stepping stone on the path to quantum computing.
According to Roese, quantum computing will revolutionize the industry in ways that surpass even the most groundbreaking advancements we have witnessed thus far. By envisioning the performance of current technology multiplied by five orders of magnitude, Roese highlights the astounding potential of quantum machine learning.
However, with great power comes great responsibility. As we approach the 2030s, maintaining cybersecurity will be paramount, as standards like AES encryption may inevitably break down. The private sector, along with numerous nation-states, is already investing heavily in proactive quantum development and adoption, recognizing the urgency to stay ahead.
In this era of unprecedented disruption, businesses must prepare themselves for the forthcoming storm of quantum computing. Roese advises organizations to use the challenges posed by generative AI as a valuable dress rehearsal for the disruptive forces that lie ahead. By navigating the choppy waters of AI, businesses can fortify their resilience and readiness to withstand the oncoming quantum storm.
Conclusion:
Businesses need to prioritize data management and understand the requirements for AI adoption. The industry must view generative AI as a stepping stone and prepare for the transformative impact of quantum computing. With careful preparation and cybersecurity measures, businesses can embrace the full potential of these technologies and navigate the upcoming quantum storm successfully.
