Essay

The idea of ressurecting the dead is a trope of much fiction and indeed one that some think might become a fact. Immortality does have some compelling features. Transhumanist and futurist Ray Kurzweil hopes to be immortalised once the technological singularity is upon us – a proposed ‘event’ where technology transcends biology, humans can upload themselves or the contents of their brains at least, and somehow a conscious being ensues.

It continues to be a fabulist’s dream if only because the notion of consciousness has never properly been defined, seen or recreated (and is referred to as the ‘hard problem’ in philosophy – not because creating consciousness is hard, but because actually explaining consciousness is surprisingly difficult for beings that have it).

Images too can appear deceptively simple. What’s not to believe about a face? We can easily identify our friends and parents from videos of them. As we all should know now though, the camera always lies – in large or small ways it feeds us seductive truths where reality is bent. Much in the same way that a political propagandist can discuss water quality improvements while the water is demonstrably unswimmable, by fudging statistics and avoiding difficult anomilies in their prose, photographs of people can amplify or avoid their darker sides, metaphorically or not.

This slippery idea of the truth was the ‘hard problem’ of straight photography and was acknowledged even soon after its birth.

The first digital cameras used in journalism brought the issue to the foreground of the general public, but in 2023, artificial intelligence applications made it much more technically easy to create realistic non-truths in imagery and the written word.

From ‘people who never existed’ (but who, thanks to scammers, often have a dating profile), to selfies with historical figures, we have been primed to question the many images put before us.

ai-based photography, or image-reconstruction if you like, has gone through a boom time over 2022-23, both overhyped and overspeculated in, but still producing unexpected and spectacular results. It’s rise to prominence is reminiscent of the rise of non-fungible tokens (NFTs) through 2021 and indeed, many of those creating nfts used early and more recent ai creation tools to quickly churn out fodder for the markets. A few experimentalists made some interesting art.

As with much technology, both nfts and the various modes of ai have often been naively used and abused with little critical consideration. Quality discussions on the pros and cons of each are hard to find, but simple-minded attacks are common, and often based on poor understandings of both the technology and the potential uses or problems.

Jon Carapiet comes to ai image generation not as a technical expert but as one who has explored the public use of images for at least 3 decades. Carapiet’s work often contains faces, but these are invariably distortions – faces pulled from the media, enlarged and showing benday dots and huge colour casts; statues that have been disfigured to the point of obliteration; faces of the well known and unknown, fragmented by pixellation and reception-snow on televisions.

Carapiet’s most recent work reminded me of those early experimental days in nft-land.

In mid-2021 on the hic-et-nunc nft platform I started seeing images of historical people, often inventors and technologists, that had been re-animated, minted and sold. These were early days in publicly available ai image generation and the videos (or animated gifs) were tiny and brief. But they held a bizarre transient pull. We knew these weren’t videos drawn from life, but the faces moved, nodded, blinked as if they’d been filmed. Some smiled, as if saying ‘yes, I acknowledge your gaze and appreciate it’. Only after a number had been presented did it become clear that the movements of body and face had a certain similarity. And the feeling of getting a deeper insight into this character was being generated in our own heads, because this person never actually moved like that.

The main nft series of these re-animations appeared to be presented as a historical window into the figures, but with an ‘isn’t it cool’ approach. Acceptance or speculation was left up to the viewer.

Carapiet felt the same bizzareness on first seeing similar re-animations on a geneology site. It seems that such sites saw the possibility to gain customers by offering re-animations of their relations. Its likely they weren’t worrying too much about whether the re-animating created untruths and false realities, as long as customers bought into the ‘cool’ compellingness of seeing those relatives ‘move’.

The images of ‘legacy’ don’t dwell on technology though, but use a photographic sensibility to query how images continue to be baffling and how *likeness* is being transformed as we encounter new photographic forms.

The ‘deep fake’ – that convincing picture or video that is in fact a construction using newly powerful algorithmic processes – has been with us for a while, taunting, with its potential for mischief and deception. It hints at the potential for ressurection, if only in memorial form.

If we imagine that we can fake the movement of a dead relative, how much more compelling is the notion of creating a whole moving and reacting persona from images and speech of that relative. No matter what some might see as ethically dubious, in 2023 companies are close to offering this, what Black Mirror showed as essentially an avatar of your departed one that can move and respond as they did while alive.

How to interrogate these seething cauldrons of ethical implications? Carapiet chose not to re-animate the dead, but rather, to de-animate the living. Having run through the ‘cool’ aspects of the technology – re-animating pre-photographic imagery such as paintings of King Henry (and of course the Mona Lisa), statues and mummified Egyptian emperors, he turned to considering the implications of using the tool on living, well-known figures.

How would they be seen by a generation that only knew them historically? The face is a signifier, a direct line to memories and for those figures who are well enough known to be recogniseable even through bad copies of photos, an instant glimpse is enough for a person to attach all they judge about the person.

Obituaries are generally accompanied by more flattering images of the person, as were the death portraits of Victorian times, to which the ‘legacy’ images have some resemblence.

In fact, Carapiet has left us with ten widely known figures, those sometimes despised or praised for their effects on modern political life. Rather than create the lie of light as Arnold Newman did with his infamous portrait of Alfred Krupps, Carapiet fed and re-fed the tool with its own outputs until figures in apparent meditation, or in a serene semblance of death were created.

These are the obituary photos of people who are not yet gone, but for whom many lines of prose will be written. Biographies will be skewed and have ommissions, reputations will be sanitised.
But for now we can gaze on and project our own thoughts onto these almost-pictures.