Pay, please: The AI, your digital replicant

The term “artificial intelligence” turns 70. The “Dartmouth Summer Research Project” took place in the summer of 1956, during which researchers discussed the development and use of artificial intelligence. As part of this conference, the US computer scientist John McCarthy proposed the term “artificial intelligence”. On the occasion of this anniversary, we have taken the liberty of republishing an older look back at the first conference and the 50th AI anniversary conference, which was first published in 2021, as “Numbers, please! Classic”. Have fun reading!

Nothing works without “artificial intelligence” (AI), at least in politics: all five parties represented in the Bundestag emphasize the important role of AI for the future of Germany in their current election programs. While this was still unthinkable during the first term of government of Chancellor Angela Merkel (CDU), in 2006 science looked back on 50 years of AI research.

On July 13, 2006, the “AI@50” conference began in Dartmouth (US state of New Hampshire) with the somewhat unwieldy full title “Dartmouth Artificial Intelligence Conference: The Next Fifty Years”. On the occasion of the anniversary, we should not only look back at 50 years of the Dartmouth Conference, but also forward – to the year 2056.

First Dartmouth Conference 1956

The first Dartmouth Conference began on June 13, 1956 at Dartmouth College and ended two and a half months later. The initiator was the LISP inventor and AI pioneer John McCarthy, who estimated the innovation leap required for human-level AI to be 1.7 Einstein and several billion. In addition to him, four other participants from the original conference that coined the term artificial intelligence took part in the celebration. Today they also died, the last survivor was Trenchard More, who left Earth in 2019.

In this respect, it is very fortunate that many of the contributions from this conference have been preserved and saved, as has the beautiful conference report (PDF) by James Moor, which appeared in the leading journal for artificial intelligence. Other locations with reports about the “Dartmouth explosion” (as Ray Kurzweil refers to nuclear weapons) are no less interesting, such as the list of science fiction (PDF) that Ray Solomonoff consumed at the time as a participant in the “Summer Research Project”.

Among the participants at the AI@50 conference were not only proven AI experts who reported on their now highly differentiated research areas, but also a philosopher. Eric Steinhart dealt with the question of the extent to which an artificial intelligence can reconstruct a personality from various digital artifacts of a human being – as a ghost or as a digital god who creates his own universe and continually programs himself.

The digital mind

“A digital ghost is an artificially intelligent program that knows everything about your life. It is an autobiography come to life. It recreates your beliefs and desires. You can survive after death as a digital ghost,” writes Steinhart. “We will be discussing a whole series of such digital ghosts over the next 50 years. As time and technology advances, there will be ever more perfect copies of their original author’s life.”

With his definition of a living AI, Steinhart was not thinking about storage spaces and web cemeteries like virtual memorials, but rather systems like DARPA’s Lifelog project or the ACM’s CARPE project, which was also discontinued. The acronym CARPE stands for “Capturing, Archiving and Retrieval of Personal Experiences”. Little remains of all these projects, even the Forever Network with thousands of life stories has disappeared and been replaced by a sports portal.

The “Personal Digital Biographer”, which AI co-founder Marvin Minsky wanted to construct as a digital shadow of himself, has not yet been realized in its entirety of data. The heise news ticker will report whether things will look different in 30 years.

(beautiful)