Predictions of the future captivate us. The end of 2024 is an excellent time to glimpse into 2025, a year that promises to be among the most interesting in technology.
Predicting the future is enchanting and engaging because it panders to our desire for control over an uncertain world. Perhaps “pander” is too strong a word, but you get the point. Predictions of the future also build hope. But here is the thing about deliberating future scenarios: They help us view, understand, and debate our choices, which is the first step toward shaping our future.
In 2024, four reports stood out in helping us understand what Artificial Intelligence (AI) will do in 2025 and beyond. The four reports include:
1. Vinod Khosla’s Plausible Tomorrows, 2035–2049
2. The Stanford Institute for Human-Centered Artificial Intelligence’s (HAI) Artificial Index Report 2024
3. Splunk’s Executive Predictions-2024, What business and technology leaders should know in the era of AI
4. WIPO’s (World Intellectual Property Organization) Patent Landscape Report for Generative Artificial Intelligence
Not all can be classified as reports. But one common thread runs through them: They all point to AI’s ability to deliver Humanity 2.0.
Vinod Khosla’s Plausible Tomorrows, 2035–2049, is a short 18-slide presentation that shines a light on a future we have all dreamed of but only hesitantly believed in. Khosla’s #1 prediction is that “Expertise will be free.” He expands on this by saying, “We will be capable of having AI-based near-free tutors for every child, and doctors for every citizen 24x7, accessible and affordable.” He then goes on to predict that labor will be near-free, thanks to robots freeing humans from the slavery of the bottom 50 percent of undesirable jobs. Among his other top predictions is that AI will play a significant role in entertainment and design, and agents will mostly provide internet access. Although Khosla does not say how this will happen, a good guess is that everything and everybody will get an API. Our agents (personal AIs) will access our data via APIs that will be self-healing, event-driven, and adaptive. This new generation of APIs will begin to emerge in 2025.
In his 2005 bestseller, The Singularity Is Near, futurist, author, and inventor Ray Kurzweil predicted that AI will surpass human intelligence by 2045, opening a period in history that will see the expansion of the mind in entirely unanticipated ways. Khosla’s predictions are on point with Kurzweil’s timeframes. Kurzweil recently wrote a sequel to his book, The Singularity is Nearer — I am putting it at the top of my reading list for 2025. (Disclosure: In 2000, Vinod Khosla, venture capitalist and co-founder of Sun Microsystems, invested in one of my internet companies).
The Stanford Institute for Human-Centered Artificial Intelligence’s (HAI) Artificial Index Report 2024 is one of the most insightful and in-depth examinations of the swiftly evolving AI landscape. It is the current Bible of AI. The report is 390+ pages long, and it is not surprising that IEEE Spectrum has put out a page extracting 15 graphs that explain the state of AI in 2024. If you do not have the time, the 15 graphs picked by Spectrum are good to drop into. These explain where investments in AI are headed, geographic leadership, who dominates the foundational models, which models are leading the race and why, their carbon footprint, etc. The last graph of the 15 provides us an insight into the attitudes of users toward AI. Does AI make you anxious and nervous? This is an excellent place to see where you stand vis-à-vis the world.
The full Artificial Index Report provides a magnificent, panoramic view of what the world of AI is dealing with. Pages 14 to 26 highlight the findings in areas such as R&D, technical performance, responsible AI, economy, science and medicine, education, policy & governance, diversity, and public opinion before diving deep into each topic. This report allows us to draw intelligent conclusions about where we are headed in 2025. For those genuinely interested in exploring the report, the raw data and charts are accessible on Google Drive.
Splunk’s Executive Predictions-2024, What business and technology leaders should know in the era of AI, is a 19-slide presentation. A Cisco company, Splunk aims to build a safer and more resilient digital world. Its #1 notice: Step changes in business impact are 12 to 24 months out. The forecasts are aimed at leadership, the C-suite, and how they will need to respond to developments around AI. For example, according to the report, resilience will become non-negotiable as AI develops. The C-suite will be delighted that the report believes organizations will focus on reducing operation complexity and cost by driving simplicity at the user level. For most C-suites, a primary challenge of AI is the growing operational complexity it is introducing into their systems. In the future, users will not have to deploy more tools to plug visibility gaps (to improve the shared context). Consolidation will lead to architectural discipline and ensure end users focus on specializing in a domain versus specific tools.
The report has Splunk CEO Gary Steel say, “In 20 years, we will have fully realized the benefits of AI in everything we do.” This forecast maps perfectly against Vinod Khosla and Ray Kurzweil’s timeframes.
WIPO’s (World Intellectual Property Organization) Patent Landscape Report for Generative Artificial Intelligence examines patent families in the core models of GenAI (generative adversarial networks or GANs, variational autoencoders or VAEs, diffusion models, autoregressive models, and decoder-based Large Language Models or LLMs) between 2014 and 2023. The report notes that over the last decade, the number of patent families in GenAI has grown from just 733 in 2014 to more than 14,000 in 2023; scientific publications have increased from just 116 in 2014 to more than 34,000 in 2023; and over 25 percent of all GenAI patents and over 45 percent of all GenAI scientific papers were published in 2023 alone. The report points out that Tencent, Ping An Insurance Group, Baidu, the Chinese Academy of Sciences, and IBM have the highest number of patents. The report provides data into the application areas for GenAI patents and insights into where GenAI is most likely to be used (spoiler: Software and Life Sciences lead this list) and the leading types of inputs and outputs used by country.
However, my interest was in understanding what the report indicated about India’s play in GenAI. Going by the numbers, we have our nose ahead: India is among the top 5 countries — after China, the US, the Republic of Korea, and Japan — where Gen AI technologies are being invented. The growth rate of GenAI patent family publications is the strongest in India (56% per year), with GAN patent families accounting for a relatively high proportion of total GenAI patents (140 GAN patents out of 1,350 Gen AI patents across five models). India is also creating above-average GenAI-related inventions in networks, with 16 patents filed in this area.
This report is an indicator of the shape of things to come — who will do what and how much in the GenAI space, and a must-read for anyone interested in understanding where to look for the first glimmer of the promised AI-driven future.
The four reports total more than 500 pages, and reading and absorbing them will take considerable time and energy. My suggestion: Use GenAI to help you find, distill, and understand what you and your industry need from these reports.
