Considerations for your next AI implementation
Currently, we are seeing two divergent approaches to AI adoption by businesses. Some organizations rush to implement solutions in an effort to generate a quick ROI, while others take a long-term view and hope for future rewards based on long-term research investments made now.
Regardless of where an organization is on its AI journey, there are common challenges they must address, including skills shortages, energy consumption, supply chain issues and budgetary constraints, with any significant AI implementation starting from £10m. It is critical that any organization considering an AI implementation invests in the right resources and technologies from the start to avoid painful and expensive headaches later.
Clear trend of investments in AI
According to figures published by Statistica, an estimated $934.2 billion was invested by companies in AI technologies between 2013 and 2022, which is steadily increasing year on year. The advent of generative AI has further boosted AI spending over the past year, with major tech companies such as Microsoft, Google and Amazon leading the way and surpassing investments from Silicon Valley venture capital firms, according to the Financial Times. Additionally, a recently published McKinsey report calls 2023 a “breakthrough year” for generative AI, with a third of respondents surveyed indicating that their organizations regularly use the technology in at least one business function.
Despite the clear trend of investment in AI, many organizations currently find large-scale AI implementations unaffordable. In addition to IT infrastructure and costs for people, it is necessary to also take into account the impact on the environment and energy consumption. Cost may be a temporary limitation for some, but organizations must have a clear path to monetization and ROI from their AI project to justify the expenditure, purchase the necessary infrastructure, and offset carbon emissions to meet regulatory requirements.
CTO International, Pure Storage.
Laying the right foundations
Regardless of the challenges, the transformative benefits and value of successful AI projects are too great to ignore. Most industries are still in the early adopter phase of AI implementation, but adoption is only increasing as use cases are defined and we move beyond the conservative thinking prevalent within many organizations. To prepare for this shift, now is the time to think about what it will take to ensure there is a solid foundation for an AI-based future.
To increase the prospects for successful AI implementation, these are the key things organizations should consider:
Accessibility of GPUs
Supply chains must be assessed and included in any AI project from the start. Access to GPUs is critical because without GPUs your AI project will not succeed. Due to the enormous demand for GPUs and the resulting lack of availability on the open market, some organizations planning AI implementations may need to look to hosting providers for access to the technology.
Power and space options for data centers
AI and its vast data sets create real challenges for already overloaded data centers, especially in the area of energy. Today’s AI deployments can require power densities of 40 to 50 kilowatts per rack – far more than the capacity of many data centers. AI is changing the networking and power requirements for data centers. It requires a much higher density of fiber, along with a larger and faster network than traditional data center providers can handle.
Energy and space efficient technologies will be crucial to successfully get your AI project off the ground. Flash-based data storage technology can help solve this problem because it is much more energy and space efficient than HDD storage and requires less cooling and maintenance than traditional hard drives. Each Watt allocated to storage reduces the number of GPUs that can be powered in the AI cluster.
Data challenges
Unlike other data-based projects that can be more selective in where the data comes from and what is taken into account, AI projects use massive data sets to train AI models and distill insights from vast amounts of information to create new stimulate innovation. . This poses significant challenges in fully understanding AI models, and how introducing new data into a model can change outcomes.
The issue of repeatability is still being grappled with, but a best practice to help understand data models and very large data sets is to introduce ‘checkpointing’, which allows models to be restored to a previous state, effectively saving time is reversed, thereby facilitating a better understanding of the data. the implications of data and parameter changes. The ethical and provenance aspects of using data from the Internet in training models have not yet been sufficiently investigated or addressed, as has the impact of (attempted) removal of selected data from an LLM or RAG vector dataset.
Investing in people
Any organization embarking on an AI journey will face skills shortages. There are simply not enough data scientists or other professionals with relevant skills available in the global workforce at the moment to meet demand. As a result, those with the right skills are difficult to come by and can command premium salaries. This will likely remain a major problem for the next five to ten years. As a result, organizations will not only need to invest heavily in talent through hiring, but also invest in training their existing workforce to develop more AI skills internally.
Conclusion
As organizations mature in their adoption of AI, develop specific use cases, refine infrastructure requirements, invest in skills, and chart a clear path to short- or long-term ROI, they may realize that the challenges can be very difficult to overcome. their own. Partnerships will be necessary for many. This is where there is a real opportunity for cloud service providers, managed service providers and other specialists to offer services and infrastructure that will help organizations realize their AI goals.
We have the best AI tools.
This article was produced as part of Ny BreakingPro’s Expert Insights channel, where we profile the best and brightest minds in today’s technology industry. The views expressed here are those of the author and do not necessarily represent those of Ny BreakingPro or Future plc. If you’re interested in contributing, you can read more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro