Artificial Intelligence (AI) continues to make headlines; this most recent round has been focused on the existential risks and the need to “slow down the pace” of AI innovation.

As with any public communication, it is important to consider the source and timing. AI is in a period of hyper innovation, with trillions of dollars being invested at a scale never previously witnessed, funding a technology that is not commonly understood by the public.

This combination creates a perfect storm for misinformation and bias (conscious and unconscious). For example, individuals and companies have staked their reputations and billions of dollars on AI, specifically the frontier AI labs, such as OpenAI and Anthropic, which are both trending towards an IPO. There is also a geopolitical debate, recognising that AI could trigger the fourth industrial revolution.

Unsurprisingly, the combination of money, politics and the mystique surrounding AI attracts mainstream media attention, with media outlets (large and small) looking to differentiate their content through increasingly provocative headlines.

I would also caution against placing too much emphasis on any individual that claims to have “insider knowledge” having worked at a frontier AI lab. Although it would be wonderful to assume everyone has altruistic goals, there will always be those few that seek to seize an opportunity for personal gain (e.g., fame, fortune).

With this in mind, I thought I would share my perspective regarding this recent debate.

However, I first want to acknowledge, even with multiple decades of experience applying emerging technologies, that I do not know how the AI revolution will unfold. In fact, I would argue that anyone who states that they do is either naive or lying.

Therefore, please consider the following words nothing more than my opinion.

First, it is not realistic to believe that we can slow the pace of AI innovation. This would require consensus from all frontier AI labs (US and OUS), alongside the open source community, which has its own economic construct.

Considering the scale of the individual and company investments, it is unlikely any party would be willing (or able) to compromise. As a result, if one party continues to accelerate their investment and the rate of innovation, the others will be forced to follow, or risk being made obsolete (which would be a disastrous position for many).

With that said, I do believe regulation will play an important role in the future of AI, helping to establish a common understanding and the rules of engagement. This would include governed policies that can be translated to technical controls, like what we have seen with other emerging technologies as they achieve mainstream adoption.

The key is to ensure that any regulation does not apply unnecessary friction, which could stifle innovation. It is important to recognise that regulation is only relevant to individuals and companies operating legitimately. Malicious actors will ignore all regulation, which could inadvertently offer them an innovation advantage.

Ironically, the best way to combat the malicious use of AI is to use AI as part of the defence. This only works if the legitimate innovation can keep pace.

Regarding the actual risks associated with AI, I would highlight four.

1. Allowing the malicious use of AI to outpace legitimate innovation. AI has dramatically lowered the barrier for the creation of malware; it also provides an ability to more effectively target (vulnerabilities) and scale (AI agent swarms) cybersecurity attacks.

2. Increased inequality, where AI becomes readily accessible to affluent demographics and countries, widening the economic gap, leading to a breakdown in social cohesion.

3. Overreliance on AI for reasoning and critical thinking with the risk of intentional bias promoting misinformation that could lead to reduced perspective or radicalisation. This can be compared to the negative impact of social networking, arguably worse due to the speed and scale information can be generated by AI.

4. System illiteracy and the rise of epistemic debt, highlighting our growing reliance on digital systems that we do not fully understand, resulting in an intellectual deficit that impacts our ability to maintain, support, and govern critical infrastructure.

I did not include the sensationalised existential risks (e.g., The Terminator, The Matrix, The Creator). I can appreciate that when you extrapolate the current pace of AI innovation, the path towards a superintelligence that could (even inadvertently) do harm is viable.

However, this is not the first time a new technology has been signalled as a planet level risk, with a recent example being the activation of the CERN Large Hadron Collider. I place confidence in the common good that these risks will be calculated and mitigated, just as they have been many times before.

Overall, I remain optimistic. Although this article focuses on the risks associated with AI, I continue to believe the benefits will be significant. The opportunity to leverage AI to solve some of the world’s biggest problems is profound, and therefore, in my opinion, a worthy and important endeavour.

artificial_intelligence ai generative_ai agentic_ai risk information_security cybersecurity software