24 November 2022
Ask the Ancestors
AI and ‘Augmented Eternity’
By Neil Tidmarsh
Our prehistoric forbears liked to keep their dead ancestors close so they could seek their advice in times of need. The bones of generations of the tribe’s leaders and wise men and women were kept in neat stacks inside the tribe’s long barrow. If there was a war or a failed harvest or a pestilence or any other crisis such as a dispute between the wives of two neolithic sporting stars, then the tribe’s shaman would unseal the long barrow and bring the remains of the appropriate ancestor out into the open and consult him or her about the best course of action.
Unfortunately, this direct access to the collective and individual wisdom and experience of our ancestors has been unavailable for a good few thousand years. It may soon be restored to us, however, according to a report from the Massachusetts Institute of Technology this week.
The MIT has launched what it calls its ‘Augmented Eternity’ project, aiming to use an individual’s digital footprint to create “a version of them that can persist even after they die” (as reported by Tom Whipple, the science editor of The Times). If the knowledge and experience which an eminent scientist or lawyer or business CEO or military general or agony aunt has gained over a lifetime could be captured and digitally stored, alongside the emotions and attitudes and psychological traits which constitute the individual’s personality, then he or she would be available for consultation for all eternity.
The idea is currently being tested on 25 people. The personal data which they leave on social media is being collected to see if it’s possible to create a digital avatar or chatbot which will “mimic exactly how an individual responds” said MIT’s Hossein Rahnama. “Might it be possible to replicate the knowledge held by experienced professionals? Can people seek legal opinions from the avatar of a famous lawyer?” he asked, adding “It was relatively easy to map those knowledge bases.” As for personality (“our intuition, our thoughts, our social behaviour, our empathy”) rather than knowledge, he said that artificial intelligence “has reached a level that allows a glimpse of these emotional legacies to be available”.
This project is clearly a huge advance on the robotic double of Taro Kono, Japan’s minister for digital transformation, reported in this column two weeks ago. So how close is this brave new world? Is AI powerful enough to identify, harvest and store such huge and subtle masses of data? Are its algorithms clever enough to analyse, access and produce the appropriate minutiae from that mass on demand? Let’s have a look at the other news items this week from the world of AI to see if they provide any kind of answer and for a glimpse of the current state of play:
(1) This week Meta launched Galactica, its new artificial intelligence search engine, claiming that it could “summarise academic papers, solve maths problems, generate Wiki articles, write scientific code, annotate molecules and proteins, and more”. IT experts weren’t impressed. One dismissed it as a “random bullshit generator”. It was taken off-line only a few days later.
(2) In Germany, Kentucky Fried Chicken invited its on-line customers to celebrate the eighty-fourth anniversary of Kristallnacht. With some fast food. On that infamous ‘Night of Broken Glass’ in 1938, massed Nazi gangs attacked Jewish targets across Germany, smashing up and burning down synagogues, shops and businesses; 90 people were killed, 30,000 were sent to concentration camps and 1400 synagogues were destroyed. “Memorial day for the Reich pogrom night!” the KFC app announced. “Go ahead and treat yourself to more soft cheese and crispy chicken. Now at KFCheese!” KFC apologised; it keeps a database of significant national events for each country in which it operates, but it seems that the algorithms which collect and select the dates are not intelligent enough to understand that not all significant events are to be celebrated or to realise that some ways of marking them are more appropriate than others.
(3) Mary McIntyre, an astronomer with 6000 followers on Twitter, posted a six-second video of a meteor on her account – and was promptly kicked off the platform because its automated moderation tools flagged the video as “intimate content” (ie pornography) which she was sharing “without the consent of the participant”.
(4) Ms McIntyre’s experience brought to light a case in the USA where Ryan Vaughan, a meteorologist, shared a video of combine harvesters working at night – which Twitter’s artificial intelligence algorithms also judged to be pornographic. They gave him the choice of deleting the video or being banned from the site.
All of this gives us some idea about how reliable any advice from a digitally recreated lawyer or scientist or CEO or agony aunt is likely to be. All in all, it looks rather as if MIT’s ‘Augmented Eternity’ and other such AI projects won’t be winning any Nobel prizes for some time to come. Not in our lifetime, at least. Which means that we won’t be able to give our descendants the benefit of our invaluable knowledge and experience for all eternity. Oh well, their loss. At least our bones will rest in peace. Unless MIT starts studying and teaching shamanism, of course. Actually, that might be a better bet for them.