AI in 2025: Growth, Challenges, and the Quest for ROI

We used Google's AI to analyze 188 predictions of what's in store for tech in 2025

Artificial intelligence has changed the way we live and work in the last two years, and going into 2025, many of those 188 reports are in agreement that AI will continue to have a huge impact. The technology will be more actively integrated into business operations across sectors, a significant number agreed. "AI was the big story of 2023 and 2024, and that has not changed. In fact, AI adoption will likely begin to accelerate in 2025 as energy and commodities companies gain confidence in use cases that promote optimization and innovation," wrote Publicis Sapient, a digital consultancy, in its 2025 outlook. But AI's use will be deployed across industries. AI is predicted to shift from a "nice-to-have" to a "must-have" tool for B2B marketers, with adoption increasing for content creation, personalization, predictive analytics, and campaign optimization," wrote EssenceMediacom, a GroupM marketing agency, in its look ahead. Banks like Barclays believe AI will play a significant role in financial markets, with investors deploying it to try to get ahead. CB Insights believes AI-powered weather prediction could transform the insurance industry in 2025. But others sound a note of caution: in its 2025 trends analysis, Zendesk highlights the risk of so-called "shadow AI" use by employees without their employers' permission, noting in some industries such shadow use has grown 250%, causing security risks. S&P Global suggests that AI, particularly generative AI, is driving a shift towards focusing on product and service quality improvements and revenue growth -- but others worry about the need to ethically develop AI, and to not assume that its training data is obtained officially.

AI will continue to grow in 2025, but it may face major challenges along the way

In 2024, artificial intelligence (AI) continued taking large and surprising steps forward. People started conversing with AI "resurrections" of the dead, using AI toothbrushes and confessing to an AI-powered Jesus. Meanwhile, OpenAI, the company behind ChatGPT, was valued at US$150 billion and claimed it was on the way to developing an advanced AI system more capable than humans. Google's AI company DeepMind made a similar claim. These are just a handful of AI milestones over the past year. They reinforce not only how huge the technology has become, but also how it is transforming a wide range of human activities. So what can we expect to happen in the world of AI in 2025? Neural scaling Neural scaling laws suggest the abilities of AI systems will increase predictably as the systems grow in size and are trained on more data. These laws have so far theorized the leap from first to second generation generative AI models such as ChatGPT. Everyday users like us experienced this as the transition from having amusing chats with chatbots to doing useful work with AI "copilots," such as drafting project proposals or summarizing emails. Recently, these scaling laws appear to have plateaued. Making AI models bigger is no longer making them more capable. The latest model from OpenAI, o1, attempts to overcome the size plateau by using more computer power to "think" about trickier problems. But this is likely to increase costs for users and does not solve fundamental problems such as hallucination. The scaling plateau is a welcome pause to the move towards building an AI system that is more capable than humans. It may allow robust regulation and global consensus to catch up. Training data Most current AI systems rely on huge amounts of data for training. However, training data has hit a wall as most high-quality sources have been exhausted. Companies are conducting trials in which they train AI systems on AI-generated datasets. This is despite a severe lack of understanding of new "synthetic biases" that can compound already biased AI. For example, in a study published earlier this year, researchers demonstrated how training with synthetic data produces models that are less accurate and disproportionately sideline underrepresented groups, despite starting with unbiased data sets. Tech companies' need for high-quality, authentic data strengthens the case for personal data ownership. This would give people much more control over their personal data, allowing them, for example, to sell it to tech companies to train AI models within appropriate policy frameworks. In 2025, Tesla intends to deploy these robots in its internal manufacturing operations with mass production for external customers in 2026. Amazon, the world's second-largest private employer, has also deployed more than 750,000 robots in its warehouse operations, including its first autonomous mobile robot that can work independently around people. Generalization -- that is, the ability to learn from datasets representing specific tasks and generalize this to other tasks -- has been the fundamental performance gap in robotics. This is now addressed by AI. For example, a company called Physical Intelligence has developed a model robot that can unload a dryer and fold clothes into a stack, despite not being explicitly trained to do so. The business case for affordable domestic robots continues to be strong, although they're still expensive to make. Automation The planned Department of Government Efficiency in the United States is also likely to drive a significant AI automation agenda in its push to reduce the number of federal agencies. This agenda is also expected to include developing a practical framework for realizing "agentic AI" in the private sector. Agentic AI refers to systems capable of performing fully independent tasks. For example, an AI agent will be able to automate your inbox, by reading, prioritizing and responding to emails, organizing meetings and following up with action items and reminders. Regulation The incoming administration of newly elected US president Donald Trump plans to wind back efforts to regulate AI, starting with the repeal of outgoing president Joe Biden's executive order on AI. This order was passed in an attempt to limit harms while promoting innovation. Trump's administration will also develop an open market policy where AI monopolies and other US industries are encouraged to drive an aggressive innovation agenda. Elsewhere, however, we will see the European Union's AI Act being enforced in 2025, starting with the ban of AI systems that pose unacceptable risks. This will be followed by the rollout of transparency obligations for generative AI models, such as OpenAI's ChatGPT, that pose systemic risks. Australia is following a risk-based approach to AI regulation, much like the EU. The proposal for ten mandatory guardrails for high-risk AI, released in September, could come into force in 2025. Workplace productivity We can expect to see workplaces continue to invest in licenses for various AI "copilot" systems, as many early trials show they may increase productivity. But this must be accompanied with regular AI literacy and fluency training to ensure the technology is used appropriately. In 2025, AI developers, consumers and regulators should be mindful of what Macquarie Dictionary dubbed the word of the year in 2024: enshittification. This is the process by which online platforms and services steadily deteriorate over time. Let's hope it doesn't happen to AI.

AI will continue to grow in 2025. But it will face major challenges along the way

La Trobe University provides funding as a member of The Conversation AU. In 2024, artificial intelligence (AI) continued taking large and surprising steps forward. People started conversing with AI "resurrections" of the dead, using AI toothbrushes and confessing to an AI-powered Jesus. Meanwhile, OpenAI, the company behind ChatGPT, was valued at US$150 billion and claimed it was on the way to developing an advanced AI system more capable than humans. Google's AI company DeepMind made a similar claim. These are just a handful of AI milestones over the past year. They reinforce not only how huge the technology has become, but also how it is transforming a wide range of human activities. So what can we expect to happen in the world of AI in 2025? Neural scaling Neural scaling laws suggest the abilities of AI systems will increase predictably as the systems grow in size and are trained on more data. These laws have so far theorised the leap from first to second generation generative AI models such as ChatGPT. Everyday users like us experienced this as the transition from having amusing chats with chatbots to doing useful work with AI "copilots", such as drafting project proposals or summarising emails. Recently, these scaling laws appear to have plateaued. Making AI models bigger is no longer making them more capable. The latest model from OpenAI, o1, attempts to overcome the size plateau by using more computer power to "think" about trickier problems. But this is likely to increase costs for users and does not solve fundamental problems such as hallucination. The scaling plateau is a welcome pause to the move towards building an AI system that is more capable than humans. It may allow robust regulation and global consensus to catch up. Training data Most current AI systems rely on huge amounts of data for training. However, training data has hit a wall as most high-quality sources have been exhausted. Companies are conducting trials in which they train AI systems on AI-generated datasets. This is despite a severe lack of understanding of new "synthetic biases" that can compound already biased AI. For example, in a study published earlier this year, researchers demonstrated how training with synthetic data produces models that are less accurate and disproportionately sideline underrepresented groups, despite starting with unbiased data sets. Tech companies' need for high-quality, authentic data strengthens the case for personal data ownership. This would give people much more control over their personal data, allowing them, for example, to sell it to tech companies to train AI models within appropriate policy frameworks. Robotics This year Tesla announced an AI-powered humanoid robot. Known as Optimus, this robot is able to perform a number of household chores. In 2025, Tesla intends to deploy these robots in its internal manufacturing operations with mass production for external customers in 2026. Amazon, the world's second-largest private employer, has also deployed more than 750,000 robots in its warehouse operations, including its first autonomous mobile robot that can work independently around people. Generalisation - that is, the ability to learn from datasets representing specific tasks and generalise this to other tasks - has been the fundamental performance gap in robotics. This is now addressed by AI. For example, a company called Physical Intelligence has developed a model robot that can unload a dryer and fold clothes into a stack, despite not being explicitly trained to do so. The business case for affordable domestic robots continues to be strong, although they're still expensive to make. Automation The planned Department of Government Efficiency in the United States is also likely to drive a significant AI automation agenda in its push to reduce the number of federal agencies. This agenda is also expected to include developing a practical framework for realising "agentic AI" in the private sector. Agentic AI refers to systems capable of performing fully independent tasks. For example, an AI agent will be able to automate your inbox, by reading, prioritising and responding to emails, organising meetings and following up with action items and reminders. Regulation The incoming administration of newly elected US president Donald Trump plans to wind back efforts to regulate AI, starting with the repeal of outgoing president Joe Biden's executive order on AI. This order was passed in an attempt to limit harms while promoting innovation. Trump's administration will also develop an open market policy where AI monopolies and other US industries are encouraged to drive an aggressive innovation agenda. Elsewhere, however, we will see the European Union's AI Act being enforced in 2025, starting with the ban of AI systems that pose unacceptable risks. This will be followed by the rollout of transparency obligations for generative AI models, such as OpenAI's ChatGPT, that pose systemic risks. Australia is following a risk-based approach to AI regulation, much like the EU. The proposal for ten mandatory guardrails for high-risk AI, released in September, could come into force in 2025. Workplace productivity We can expect to see workplaces continue to invest in licenses for various AI "copilot" systems, as many early trials show they may increase productivity. But this must be accompanied with regular AI literacy and fluency training to ensure the technology is used appropriately. In 2025, AI developers, consumers and regulators should be mindful of what Macquarie Dictionary dubbed the word of the year in 2024: enshittification. This is the process by which online platforms and services steadily deteriorate over time. Let's hope it doesn't happen to AI.

ROI and more regulations: Here's what to expect from AI in 2025

As the AI momentum continues into 2025, we look at some expert predictions for the tech in the coming year. It wasn't hard for experts to predict that artificial intelligence (AI) technologies would continue to gain momentum in 2024. However, the sheer importance that AI - especially generative AI (GenAI) and widely available large language models (LLM) - have garnered since exploding onto the scene just a few years ago is still astounding. While this momentum isn't looking to slow down in 2025, expert analysts are now saying that business leaders will focus on getting their return on investment (ROI) on AI by putting the lessons they learned over the years into practice. Dr Marc Warner, the CEO of AI consulting firm Faculty, says that senior leaders need to stop viewing AI as "experimental", and need to start treating it as an essential part of business transformation. He also advocates for a business strategy that integrates AI, rather than introducing standalone strategies, which will, according to him, provide the most success. "Introducing new goals driven solely by AI has the potential to overcomplicate business processes and distract from solving key challenges," he says. Moreover, newer developments on the regulatory side of things, including the European Union's landmark AI Act, the UK's recent attempt to regulate data usage by AI firms as well as the US's flip-flopping stance around the technology could succeed in containing some of the 'Wild-West' attitude of AI innovators, compelling them to fall in line. Beware of premature of AI roll-backs According to Forrester's 2025 predictions, impatience with AI ROI "could prompt enterprises to prematurely scale back investment". Warning against this, analysts from the market research company say that businesses are achieving improved customer experience and productivity with AI, however, ROI could still take longer than decision-makers anticipate. According to its analysis, 49pc of US GenAI decision-makers said that their organisation expects ROI from AI investments within one to three years, while 44pc said three to five years for the same. Meanwhile, a Basware report from last month said that 48pc of senior finance leaders would be wary of further investments into AI if their initial investments don't deliver within 12 months. However, Nitesh Bansak, the CEO of digital product engineering company R Systems, said that more and more organisations will take a practical approach towards what they can achieve using AI in a short period of time. "AI has had a lot of hype - and that will continue - but more and more tech leaders will place a higher emphasis on ensuring measurable ROI - especially what can be achieved in the same fiscal year or under 12 months," he says. "For example, if an organisation requires complex coding to enhance a new product, using AI to automate routine development tasks can free up needed time and gain an immediate benefit." TuringBots will accelerate SDLC This leads into another industry prediction for the year; more businesses will infuse AI development tools to assist their development teams. TuringBots are AI development tools that automate and assist development teams, and 24pc of executive-level respondents told Forrester that they will leverage AI and GenAI across the software development life cycle (SDLC) within the next year. And Bansak agrees. "In the coming years, generative AI copilots, equipped with expanded memory and an agentic mesh framework, will increasingly enhance SDLC," he said. "With the agentic mesh, enterprises can deploy AI copilots that offer services such as high-fidelity UI testing, dynamic UX design, advanced prompting and domain-specific customisation. "This interconnected intelligence will automate routine tasks, allowing engineers to focus on high-value, innovation-driven initiatives," he adds. The vitality of 'clean data' Much of the conversation around AI data usage this year has been in relation to the origin and quality of data used by firms. Transparency around data usage and a rethinking around future regulation of AI is key to protect businesses and content owners from copyright infringement, Sahara AI's CEO told SiliconRepublic.com last month. "Good, clean data is imperative for enterprises," Bansak says. "If a company is leveraging AI in a chatbot feature, they must consider what data is being used to train the generative AI model and ask critical questions. Where did the model obtain its data? What kind of data is included? Has the data been evaluated and vetted to ensure its accuracy? "Poor quality, inaccurate, or incomplete data can cause multiple issues in AI training and output, ultimately negating the benefits that the AI was initially meant to create." This past week, the UK government launched a consultation with the intention of introducing regulations around AI and copyright. While its proposals have received mixed reactions from stakeholders, regulations are just around the corner and AI firms need to be wary of what data they use. "AI will be driven by governmental entities and there will be a more focus on the regulation of AI," says Marais Bester, a consultant at SHL, a data analytics and consulting firm. "There will be a focus on how we utilise the tools to ensure that people's confidentiality is protected," he adds, "but also in an ethical way so that AI doesn't get too smart before we lose control of the technology." However, while Europe is moving towards stricter regulations, the US's stance on AI remains to be seen. Incoming president Donald Trump's first major AI policy move would likely be to repeal outgoing president Joe Biden's executive order on AI, which sought to address the developing technology's threat to civil rights, privacy and national security, Time Magazine opines. Data centres will take centre stage According to Sure Valley Ventures, the AI-focused venture capital firm operating in Ireland and the UK, global data centre power consumption is expected to double, potentially reaching 1,000TWh (terawatt hours) by 2026. Moreover, newer technological breakthroughs like high-density racks of 30-40kW are expected to become the norm. In order to supplement the growing need for energy, the International Energy Agency expects that nearly 25pc of data centre power will come from renewable energy sources by the end of 2025. However, while the rate of renewables are expected to rise, the meteoric rise in energy requirements will ultimately result in a rise in overall energy consumption, a significant portion of which is still fossil-fuel based. For example, earlier this year, Google signed an agreement with Kairos Power to purchase nuclear energy, claiming that the deal will "accelerate the clean energy transition across the US". However, the software giant, which has goals to reach net-zero carbon emissions by the end of the decade, also recorded a carbon emissions growth of nearly 50pc compared to its 2019 levels. Don't miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic's digest of need-to-know sci-tech news.

ROI from AI and more regulations, here's what to expect from 2025

As the AI momentum continues into 2025, experts warn leaders to place expectations on returns. It wasn't hard for experts to predict that artificial intelligence (AI) technologies would continue to gain momentum in 2024. However, the sheer importance that AI - especially GenAI and widely available large language models (LLM) have garnered since exploding onto the scene just a few years ago, driving conversations around new regulations, is still astounding. While this momentum isn't looking to slow down in 2025, expert analysts are now saying that business leaders will focus on getting their return on investment (ROI) on AI by putting the lessons they learned over the years into practice. Dr Marc Warner, the CEO of Faculty, an AI consulting firm says that senior leaders need to stop viewing AI as "experimental", and start treating it as an essential part of business transformation. He also advocates for a business strategy that integrates AI, rather than introducing standalone strategies, which will, according to him, provide the most success. "Introducing new goals driven solely by AI has the potential to overcomplicate business processes and distract from solving key challenges," he says. Moreover, newer developments on the regulatory side of things, including the European Union's landmark AI Act slated to come into force - in parts - throughout the coming years, the UK's recent attempt to regulate data usage by AI firms as well as the US's flip-flopping stance around the technology could succeed in containing some of the 'Wild-West' attitude of AI innovators, compelling them to fall in line. Beware of premature of AI roll-backs According to the Forrester's 2025 predictions, impatience with AI ROI "could prompt enterprises to prematurely scale back investment". Warning against this, analysts from the market research company say that businesses are achieving improved customer experience and productivity with AI, however, ROI could still take longer than decision-makers anticipate. According to its analysis, 49pc of US generative AI decision-makers said that their organisation expects ROI from AI investments within one to three years, while 44pc said three to five years for the same. Meanwhile, a Basware report from last month said that 48pc of senior finance leaders would be wary of further investments into AI if their initial investments don't deliver within 12 months. However, Nitesh Bansak, the CEO of R Systems, a digital product engineering company, said that more and more organisations will take a practical approach towards what they can achieve using AI in a short period of time. "AI has had a lot of hype - and that will continue - but more and more tech leaders will place a higher emphasis on ensuring measurable ROI - especially what can be achieved in the same fiscal year or under 12 months," he says. "For example, if an organisation requires complex coding to enhance a new product, using AI to automate routine development tasks can free up needed time and giving an immediate benefit." TuringBots will accelerate SDLC This leads into another industry prediction for the year; More businesses will infuse AI development tools to assist their development teams. TuringBots are AI development tools that automate and assist development teams, and 24pc executive-level respondents told Forrester that they will leverage AI and GenAI across the software development lifecycle (SDLC) within the next year. And Bansak agrees. He says that "in the coming years, generative AI copilots, equipped with expanded memory and an agentic mesh framework, will increasingly enhance SDLC. "With the agentic mesh, enterprises can deploy AI copilots that offer services such as high-fidelity UI testing, dynamic UX design, advanced prompting, and domain-specific customisation. "This interconnected intelligence will automate routine tasks, allowing engineers to focus on high-value, innovation-driven initiatives," he adds. The vitality of 'clean data' Much of the conversation around AI data usage this year has been in relation to the origin and quality of data used by firms. Transparency around data usage and a rethinking around future regulation of AI is key to protect businesses and content owners from copyright infringement, Sahara AI's CEO told SiliconRepublic.com last month. "Good, clean data is imperative for enterprises," Bansak says. "If a company is leveraging AI in a chatbot feature, they must consider what data is being used to train the generative AI model and ask critical questions. Where did the model obtain its data? What kind of data is included? Has the data been evaluated and vetted to ensure its accuracy? He explains that "poor quality, inaccurate, or incomplete data can cause multiple issues in AI training and output, ultimately negating the benefits that the AI was initially meant to create." This past week, the UK government launched a consultation with the intention of introducing regulations around AI and copyright. While its proposals have received mixed reactions from stakeholders, with regulations just around the corner, AI firms need to be wary what data they use. "AI will be driven by governmental entities and there will be a more focus on the regulation of AI," says Marais Bester, a consultant at SHL, a data analytics and consulting firm, who adds that "there will be a focus on how we utilise the tools to ensure that people's confidentiality is protected, but also in an ethical way so that AI doesn't get too smart before we lose control of the technology." However, while Europe is moving towards stricter regulations, the US's stance on AI remains to be seen. Incoming president Donald Trump's first major AI policy move would likely be to repeal outgoing president Joe Biden's executive order on AI, which sought to address the developing technology's threat to civil rights, privacy and national security, Time Magazine opines. Data centres will take centre stage According to Sure Valley ventures, the AI-focused venture capital firm operating in Ireland and the UK, global data centre power consumption is expected to double, potentially reaching 1,000 TWh (terawatt hours) by 2026. Moreover, newer technological breakthroughs like high-density racks of 30-40kW are expected to become the norm. In order to supplement the growing need for energy, the International Energy Agency expects that nearly 25pc of data centre power will come from renewable energy sources by the end of 2025. However, while the rate of renewables are expected to rise, the meteoric rise in the requirement of energy will ultimately result in a rise in overall energy consumption, a significant portion of which is still fossil-fuel based. For example, earlier this year, Google signed an agreement with Kairos Power with the aim of transitioning towards clean energy by 2030. However, the software giant, which has goals to reach 'net-zero' carbon emissions by the end of the decade, also recorded a carbon emissions growth of nearly 50pc compared to its 2019 levels. Don't miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic's digest of need-to-know sci-tech news.