The desire for safety, for oneself and one’s loved ones, is deep-rooted in the human psyche. We go to great lengths to seek appropriate protection. The contemporary landscape reminds all too clearly of the impacts of this being brutally removed. Look at the Ukraine and Gaza for two horrific cases in point.
Profound and shocking though such traumas are, one doesn’t need to be caught up in armed conflict to be robbed of one’s safety. The insidious attacks on personal safety through, for example, domestic or psychological abuse can be deeply undermining of personal autonomy and wellbeing. In some ways, the closer to home the intrusion, the more vulnerable one feels.
Of course, our traditional definition of ‘home’ as a metaphor for our safe space has become significantly blurred by the intrusion of the virtual world. It is no longer easy to define the boundaries of our operating space. This diffusion of the virtual domain into the real opens countless opportunities and enables an expansiveness hitherto impossible, except perhaps for the privileged few. However, simultaneously this diffusion enables an oft unperceived and unwanted intrusion. Our domain is compromised. Our safety is violated.
This argument is well-understood and has prompted many – if not enough – important initiatives to support online safety. The draft Online Safety Bill introduced by the UK Government is a recent example. It is undoubtedly well-intentioned but the wide range of critiques of the legislation evidences the complexity of balancing our freedom and protection in these ambiguous hinterlands.
The threat, however, is now much wider and less visible than traditional cyber intrusions. Privacy considerations are manifest and often subtle. Cyber-stalking is an obvious, and profoundly disturbing, example. Less obvious is the commercialisation by others of our private data for reward. Sometimes this is blatantly done without our permission. Oftentimes it is claimed to be done with our permission because we ‘consented’ in the dense, multi-page T&Cs that none of us ever read in their entirety. And, surreptitiously, unsighted in the background, there are ‘bots’ scraping our data and using it, unbeknown to us, for whatever purpose.
In her recent, much-applauded book Privacy is Power, Prof Carissa Véliz argues powerfully for the end to the data economy where such abuses of our privacy to a greater or lesser extent expose our personal vulnerability.
There is an additional dimension whose impact is perhaps yet to be truly felt but the ubiquity of which will soon impact our safe domain, and that is artificial intelligence. The proliferation of AI deeply embedded in a range of services we regularly procure and consume exerts its influence unseen deep within our perceived safe space.
Of course, at one level it could be argued that this is mere internal detail; it is simply a technical vehicle for efficient delivery of services. However, at a deeper level, when we recognise that current approaches to AI are built upon massive datasets scraped from the cybersphere – often unseen and without permission – and, furthermore, that such ‘intelligence’ is ‘owned’ by massive corporates who thereby exert massive control over ever-wider aspects of our lives, the scale of the potential impact emerges. Another form of intrusion!
Moreover, the internal construction of these technologies currently introduces inherent biases into the decision-making mediated by the AI which means that the intrusions are differential. In other words, to misuse a phrase from George Orwell, all of us are vulnerable, but some of us are more vulnerable than others.
We need to care about this, and we need to care about it now.
The recent promotion by the UK Government of the topic of AI safety, culminating in the first global AI Safety Summit at Bletchley Park In November 2023 is greatly to be welcomed. But this is just the start. The issues of ethics and trust must be core considerations of AI and all information technologies going forward.
Safety is unquestionably an inalienable human right. Technology can do so much to engender it – and so much to threaten it. Trust must ever be a central consideration for policy makers and technologists worldwide.
