Advances in artificial intelligence (AI) continue to happen at an extraordinary pace. News of breakthroughs in machine learning, computer vision, data science, and machine-human interaction come almost daily. Growth is massive, and it is impacting all sectors.
According to a new forecast from the technology research company Gartner, worldwide artificial intelligence (AI) software revenue is forecast to total $62.5 billion in 2022, an increase of 21.3% from 2021.
The AI boom has been fueled by incredible technological leaps, yet with progress comes challenges. We asked five global AI experts to weigh in with their thoughts on trends, breakthroughs, and challenges in the year ahead.
Adji Bousso Dieng is a computer science professor at Princeton University and a research scientist at Google AI; her work spans probabilistic graphical modelling, statistics, and deep learning.
Dieng flagged up the growing trend of applying AI to "critical domains" that impact everyone in society, such as climate and biology. "I am expecting AI to continue to break ground in more scientific domains in 2022 and beyond, and I am very excited to live in this era where we are able to tackle important problems leveraging data, computation, and human insights," she said.
According to Dieng, AI research has tended to focus on building increasingly large models, supported by investment from both industry and academia. She expects this trend continue, but adds, "I personally hope we will shift energy and focus away from chasing bigger and bigger models and focus on tackling the important problems facing humanity."
Daniela Rus is a professor of electrical engineering and computer science at Massachusetts Institute of Technology (MIT), where she also serves as director of the university's Computer Science and Artificial Intelligence Laboratory (CSAIL). As an expert in AI, robotics, and data science, Rus said she has noticed a "heightened awareness about the challenges with today's AI solutions" and expects this to inform future trends.
Among Rus' predictions for 2022 are developments in data availability, interpretability, and privacy. She pointed out that the massive datasets that deep neural networks depend upon need to be manually labeled, and are not easily obtained in every field. "The quality of that data needs to be very high, and if the data is biased or bad, the performance of the systems trained on this data will be equally bad."
Rus said AI currently faces significant challenges around privacy and trust. "As we gather more data to feed into AI systems, the risks to privacy will grow. So will the opportunities for authoritarian governments to leverage these tools to curtail freedom and democracy in countries around the world."
Tackling misinformation is also a pressing concern, said Rus, "As deepfakes get better and more widespread, this problem will become more urgent for national security."
All these issues increasingly will fuel AI advancement, "We know these problems are coming and the research, business, and policy communities are already working on solutions," Rus said.
Concerns around misinformation and the abuse of AI also inform the predictions of Joanna J. Bryson, a professor of ethics and technology at the Hertie School, a private university in Berlin, Germany. Bryson's research focuses on the impact of technology on human cooperation, and AI/ICT (Information and Communications Technology) governance.
According to Bryson, AI breakthroughs and challenges are ultimately tied to questions of digital governance. "Are we using technology in a way that is safe, just, and equitable? Are we helping citizens, residents, and employees flourish?" she asked.
In the year ahead, Bryson expects cybersecurity and misinformation to be hot topics in AI. She flags up the situation in Russia and the Ukraine, and the impact of AI on the U.S. election and on democracy in individual states. "I am also wondering whether this will be the year with France heading the EU and Germany the G7 when we finally make progress on engaging with transnational antitrust," she said.
Data use, trust, and privacy are not the only societal concerns fueling AI advances. Climate change also features prominently on experts' lists.
Rus expects that in 2022, "We will understand more clearly the carbon footprint of machine learning and get more serious about Sustainable AI." She pointed out that while AI can help slow the impact of global warming in areas like energy efficiency, food security, deforestation, and biodiversity perseveration, AI systems themselves consume enormous amounts of energy.
"We need to develop simpler, yet more powerful, interpretable and causal models. This, in turn, can drastically reduce the carbon footprint of AI," Rus said.
Murray Shanahan, a professor of cognitive robotics at Imperial College London in the U.K., a senior research scientist at DeepMind, and author of The Technological Singularity, also pointed to efficiency and sustainability as driving forces in the year ahead.
"One trend I hope we'll see is machine learning systems that can do more with less, that can learn to generalize better with less data, fewer parameters, and less computation and therefore using less energy," said Shanahan.
Shanahan suggests pretraining models that can be re-used or fine-tuned, and learning better abstractions, are good steps toward doing more with less, "In general, we want systems that can distill as much as possible from the data they've previously seen for later re-use. Humans are very good at this, and our current machine learning systems have a way to go to catch up," he said.
Vineeth N. Balasubramanian, a professor of computer science and engineering at the Indian Institute of Technology in Hyderabad, also predicts trends motivated by issues like climate change and trust.
"As AI gets increasingly used for environmental conservation and renewable energy technologies, AI technologies themselves are creating the next environmental risk in the datacenters required to support AI technologies," he said.
Balasubramanianwhose research focuses on machine learning and computer visionexpects we will see a move towards more responsible, explainable, and trustworthy AI this year. In practical terms, this means more streamlining, formalizing, and evaluating of AI.
He anticipates there will be additional efforts to create AI models that address multiple tasks simultaneously. They could take the form of "Multimodal AI that brings together vision, language and speech technologies," Balasubramanian said, as well as more rapid AI development, prototyping and deployment through no-code/low-code AI, and greater success in transferring AI models from one domain/task to another.
Balasubramanian also raises AI accessibility as a pressing concern. "Most AI applications are targeted towards the first billion on the planet, be it developed nations or urban communities in other nations," he said, adding that additional efforts are needed "to highlight the needs of the 2nd and 3rd billion, and their needs may gradually gain importance."
According to Rus, while advances in automation and robotics have the potential to make people's lives easier, these technologies also can displace human workers. "We will also see focused efforts to anticipate and respond to the economic inequality this could create, in the form of advancements in intuitive interactions between humans and machines, with machines adapting to humans rather than the other way around."
Rus also expects the impact of AI breakthroughs to filter through to real-world scenarios and solutions, such as the development of new education programs aimed at young students and reskilling programs aimed at the current workforce.
Dieng is also thinking about the effects of building machines that ape human activity. She pointed out there has been a narrow focus within the AI community of "building models and algorithms to accomplish tasks as well as humans." This, she explained, reflects the pursuit of artificial general intelligence (AGI) and has led to the current trend for developing increasingly large models.
"I think this narrow focus on building models to imitate what humans can do is a challenge in the community, in that it constitutes a missed opportunity for applying AI to critical domains affecting us all."
However, Dieng is optimistic about the future, "Fortunately, there are researchers working at the intersection of AI and those domains and I am hoping they will receive a lot of funding support and that their work will receive more attention."
That 2022 will see technological advances in data science, machine learning, computer vision, and other technologies is inevitable. However, as AI becomes embedded in society and more directly impacts peoples' lives, future trends look set to be increasingly driven by the consequences of such a seismic shift.
Karen Emslie is a location-independent freelance journalist and essayist.
No entries found