"Very few businesses are planning for superintelligence": Preparing for post-human work
If businesses can't control the AI available today, how can they hope to lock down a machine smarter than our species?
AI superintelligence is moving out of the realm of science fiction and becoming an increasingly real-world capability every day. As we watch it evolve in real time, the concept of technology outperforming humans across most or all cognitive tasks no longer seems as far-fetched as it might have done this time last year.
We are even seeing warnings from senior executives at companies at the forefront of AI development, whilst the dangers of AI are being recognized across all levels of government.
The debate reached the House of Lords earlier this year, and, more recently, MPs have addressed the potential need for kill-switches. Although some experts believe AI could exceed human capabilities by 2030, the general consensus is that AI superintelligence is still several years, or even decades, away.
Whether or not this 2030 prediction comes true, businesses are facing a more immediate dilemma in harnessing the power of the AI technologies available today, let alone future systems that are far more advanced.
Looking beyond the hype
Very few businesses, if any, are well placed to plan for the emergence of superintelligence. Rather than a lack of interest, the biggest barrier to further progress on this front is the current lack of compute power available.
Building increasingly capable AI models demands vast processing capacity, which has triggered an unprecedented wave of investment in hyperscale data centers and supporting infrastructure around the world.
This highlights the practical realities of scaling AI. Advanced data centres require enormous amounts of electricity, while many rely on significant volumes of water for cooling.
READ MORE: Shadow AI is becoming a major systemic risk. Even intelligence agencies are getting worried
Proposed developments are already leading to widespread community concerns, and authorities are having to weigh the economic benefits of AI infrastructure against environmental impact and pressure on local resources.
One proposed solution to these challenges is quantum computing. This technology has the potential to change that equation by providing vastly greater processing capabilities.
However, experts agree there is a lot more to be done before this technology is viable at the required levels, pushing the realities of scaling AI superintelligence further down the road.
Understanding the risks
That said, we can all agree that AI moves fast. Since the launch of ChatGPT in 2022, it has become deeply ingrained in our daily lives and continues to evolve at unprecedented speed.
Businesses have also experienced this shift, with many rushing to deploy the latest technology over the last few years. However, without clear integration strategies, they will struggle to optimize their current solutions, so the outcry about the dangers of AI is somewhat premature.
For instance, many businesses still lack a clear understanding of what information AI systems can access, how it is being used, and where it is ultimately shared. The issue is often not a lack of AI capability, but a lack of visibility and control, particularly as organizations are looking to apply it at scale.
This is where businesses must be aware of the risks that accompany new technology. As AI models consume increasing volumes of data, uncertainty around ownership and privacy becomes harder to manage. The hype around AI makes establishing the right foundations seem an unglamorous task.
READ MORE: TalkTalk collapse could pose a "genuine risk to life", Westminster warns
]But measures such as data loss prevention and information classification are increasingly important as companies continue to adopt a wider range of AI solutions. Without stringent controls, sensitive information is left vulnerable and at risk.
On top of this, businesses have to contend with the risks that arise from shadow AI. This creates additional complexity as companies are unable to maintain constant oversight into how corporate information is used.
Then there are the governance issues to address. Businesses need to decide which use cases should be entrusted to AI and where human oversight should remain essential.
Human-in-the-loop approaches cannot be discounted while risks around reliability and potential hallucinations are still a concern, particularly in environments where decisions have significant consequences, such as finance, cybersecurity or HR.
These issues have historically proven difficult for businesses to address at scale, and the results rarely deliver immediate commercial benefits. As a result, data readiness has often been treated as a lower priority than more visible AI initiatives. This is a mistake companies cannot afford to make. Any initiative without the correct foundations in place is unlikely to succeed.
Building a solid foundation
As it stands, the biggest divide we are seeing is not between humans and AI systems, but between businesses that successfully integrate the technology and those that struggle to move beyond experimentation.
Starting the groundwork now will mean that, when AI superintelligence does arrive, it will be in a world better prepared for it following years of incremental change and adaptation.
The companies that take this approach will be best positioned for a superintelligent future, not those who spend time trying to predict when it will arrive.
Jay Fitzhenry is vCTO for AI at Node4
