AS 2023 draws to a close, Designit’s Miguel Sabel, global director of strategy and sustainability predicts six major trends and moments regarding the world of AI as we head into 2024:
1. The AI arms race will turn into guerrilla warfare
“The intense competitive and even geo-politics power struggle that we witnessed throughout 2023 is set to persist into 2024. Amid this fierce competition, through internal information leaks, we have gained insights into how industry-leading companies acknowledge the challenges of maintaining a sustainable competitive advantage solely through their models. This difficulty arises not just from the threat posed by rival corporations, but also due to the proliferation of open-source models.
“In the realm of product and user experience, OpenAI has already pioneered the path for crowdsourced innovation with the rise of user-customised GPTs. A marketplace for these innovations is currently under construction, with the belief that “the best GPTs will be crowdsourced by the community.”
“Crowdsourced innovation can be a way to achieve more diverse, contextualized, and problem-specific products. These innovators will enter a highly competitive space that counts investments by billions, but why not be optimistic? How many knew about OpenAI at the end of 2022 when ChatGPT was released?”
2. Regulation
“In 2023, we’ve witnessed a significant surge in government initiatives concerning AI: the White House Executive Order, the UK AI Safety Summit, and the implementation of the EU’s AI Directive. The year also marked a notable increase in international cooperation, exemplified by the outcomes of the G7 Hiroshima AI Process, which resulted in 11 International Guiding Principles to govern AI.
“However, these efforts have not been without criticism. Critics would argue that they are primarily driven by geopolitical power fights, heavily influenced by interested doomsayers, and perhaps, insufficiently addressing the current harms to ordinary citizens.
“Looking ahead to 2024, let’s take a different approach – viewing these constraints not as impediments, but as catalysts for creativity. This mirrors the impact of sustainability regulations, which, despite presenting challenges, have also served as incentives. If AI follows the same path as sustainability 2024 will see a huge uplift in the development of new businesses, enriched experiences, and likely, see the groundwork for future competitive advantages being put in place.”
3. Accelerated maturation
“As the ‘magic dust’ begins to settle during 2024, GenAI will be expected to reach the lofty expectations that have been laid upon it.
“Researchers and the industry will need to address its fundamental limitations. Fundamental because some are hygiene factors, and also because some seem to be inherent to GenAI current form. Issues such as copyright infringement, biases, hallucinations, and lack of traceability are prime examples.
“Moreover, GenAI will be expected to start delivering on promises of billions of dollars in efficiency gains and the liberation of thousands of human hours for higher-value activities. That’s not a small feat to achieve.
“I would like to see progress on these fronts in 2024 and also advocate for the development of a critical perspective that collectively helps us set future expectations. One that shift us from the “move fast and break things” paradigm to a new ethos of “move fast and do better.”
4. Disappearing computing / Headless interaction
“For years, technology has promised ubiquitous, transparent artificial intelligence dedicated to human augmentation. This concept, once confined to the realm of science fiction, has begun to materialise in the form of voice assistants, which have found a place in many homes with varying degrees of success. Upcoming products like the Humane AI pin continue to fuel these expectations.
“While technology is likely on the verge of making this possible, we have yet to witness a product based on headless interaction that truly meets these expectations. 2024 could be the year in which we do see significant progress though – both in terms of user experience and, more fundamentally, in the creation of trust.
“Firstly, we can observe how GenAI is defining new experience models. Its novelty requires new interaction paradigms – consider recent developments like co-pilot technology, spatial computing, or ephemeral apps, which were virtually unheard of not long ago. These advancements will bring us closer.
“Perhaps even more crucially, Gen AI is amplifying existing concerns about responsible technology. This is not only a public concern but also a significant issue within the very companies that design and build these products. If we are to integrate an invisible agent into our lives, it is imperative that we trust it.”
5. Adoption tensions as opportunities for innovation
“We find ourselves amidst a significant hype cycle, yet the exact position within the curve remains uncertain. Each new release stirs excitement, yet there’s a growing concern about a potential bubble burst, especially when experts identify a specific decline in the quality of responses of AI tools like ChatGPT due to model changes.
“In this fast-paced and uncertain environment, AI adoption is not uniform. It’s limited to a select group, primarily due to disparities in access to necessary tools, hardware, and a comprehensive understanding of its capabilities.
“This tensioned adoption scenario is likely to persist, even as AI-enabled products and services become more integrated and commonplace. The relentless pace at which corporate products are released only reinforces this.
“But there’s hope. Outsiders can bring in 2024 change through leapfrogging innovation, especially in areas or among groups that haven’t been part of this tech wave yet. We’ve seen before how this is the way to create new products, services and experiences that quickly become the norm for everyone. That would be true disruption.”
6. Expectations about the models
“The GenAI explosion is being driven by surprising and unexpected models. If anything – prediction is futile, as trying to anticipate the next trick of these GenAI models is impossible. But there are some things we can expect.
“Only recently, they have become more connected and able to handle near real-time data, and we expect this to extend in 2024. Data pollution or even malicious data injection might become a problem that will become increasingly hard to counter. The opportunities for innovation will expand too.
“These safety issues are exacerbated by the fact that today, GenAI models are de facto black boxes which probably are impossible to decipher. We have hopes about this being challenged through the development of Explainable AI or XAI. The goal of XAI is to make the decision-making processes of AI and machine learning models more comprehensible, thereby enabling users to understand, trust, and effectively manage these technologies. How can this be mitigated with current AI technologies? Is it even possible?”