ARTIFICIAL INTELLIGENCE (AI) & INTELLIGENCE AUGMENTATION (IA)

ARTIFICIAL INTELLIGENCE (AI)

Artificial Intelligence, or AI, as its name implies, is a human-made form of intelligence capable of competing or even surpassing that of our own. AI can substitute, supplement and/or amplify practically all manual tasks performed by humans, utilizing machines or robots, and has done so automating a great number of tasks.

Today, it is common for large factories to operate with practically no workers and a few white-collar employees behind computers screens supervising the entire factory operations. AI can also substitute, supplement and/or amplify a growing number of mental tasks, until recently the sole province of humans, using sophisticated computer algorithms. The major advantage of AI is that it can learn on its own, by means of trial and error, once the rules have been set and objectives have been specified. This is done by discovering patterns and/or relationships through the analysis of massive amounts of data and then applying these patterns/relationships to improve human decision-making.

However, there are still quite a few AI-related challenges:

  1. AI may lead to large-scale unemployment, or at least massive employment displacement, as AI is capable of performing a growing number of mental tasks more effectively and efficiently than humans.

  2. AI can produce huge wealth inequalities, as those exploiting AI can reap substantial benefits, where the rest may be left behind.

  3. There are fears that in the future AI could become smarter than humans. This would mean taking over practically all tasks and decisions done by people, with serious implications for the future role of humans in the economy and society and maybe for humanity itself.

While AI is considered by many as a competitor to our own, human intelligence, there are at least two facts that contradict this view. First, AI is still a long way from approaching human intelligence and becoming a danger. Second, AI is capable of contributing significantly in augmenting our own intelligence, thus allowing us to keep up with AI advances. Human and artificial intelligences are complementary rather than adversarial.

AI is strong in three aspects and weak in all others where human intelligence excels. At present AI is superb in games where it easily beats world champions and excellent in image and speech recognition where it has reached and, in many cases, has surpassed humans. For instance, AI algorithms can recognize cancer in X-Rays as well or better than specialized doctors can while other algorithms can now understand spoken language, speak fluently and translate between languages on par with humans.

Apart from the above three breakthrough achievements, AI has not shown any other concrete accomplishments at least until now. AI cannot deal with ambiguity, is unable to understand content, is incapable of acquiring common sense and cannot be creative and entrepreneurial, areas where humans naturally excel. What is needed, therefore, is AI to be used to complement and augment our own intelligence in what is being called Intelligence Augmentation (IA) that will allow us to keep up with advances in AI, thus exploiting its advantages while avoiding its drawbacks.

INTELLIGENCE AUGMENTATION (IA)

Intelligence Augmentation (IA) is a way to amplify our own intelligence by exploiting advanced AI. Consider a word processor that is developed in an adversarial position to human writers. There is no doubt that such processor will be superior in spelling to practically all humans, but this means nothing since correct spelling requires searching a big dictionary. Nonetheless, if the processor can indicate that a word is misspelled or better figure out and automatically correct the misspelled word, then we talk about IA. Such a word processor, powered by AI, would allow us to concentrate on the more creative aspects of writing, such as making the story more exciting, having more time to express the writer’s thoughts or use the thesaurus provided by the processor to find appropriate synonyms. Finally, the processor can go a step further by suggesting the next word in the text and providing indications that certain phrases may be grammatically incorrect. Such help does not prove the supremacy of the word processor, it simply exploits the computer’s ability to certain tasks that we, humans are not good at, while allowing the writer to concentrate on the task that humans can excel, that of writing an interesting and creative story.

The value of IA can become more evident in games like Chess or GO where the concept of Centaur is further advanced. For a period of time the world chess champion was neither a computer nor a human but a human using computer(s). This has been called the “centaur” player (half human/half computer), like the mythical centaur, half-human/half-horse. In such playing, the human is in charge of strategy and the computer responsible for the tactics and for checking the correctness of possible moves. Such centaur played by Steve Crampton and Zackary Stephen won the chess championship in 2005 beating both grandmasters and tough computer programs. Centaur programs are still doing well in championship even though computers have become faster and AI algorithms have advanced considerably easily beating chess grandmasters. Still chess grandmasters train on computers to perfect their games with the consequences of improving the level of their games and resulting in more tie matches. Finally, chess and GO programs sometimes play highly creative, unusual moves, surprising grandmasters who, however, after studying them they incorporate them in their games, increasing the power of their games and creating more exciting ones.

Accordingly, future developments must focus on turning AI into IA, working toward a symbiosis of humans and machines to complement and support each other rather than being considered as adversaries in a lose/lose competition. Such symbiosis will augment human decision-making abilities and will significantly reduce the perceived dangers that AI will end human supremacy.

The purpose of IFF is to keep abreast of new advances in AI and IA, provide such knowledge and its implications for teaching and research to all departments of UNIC, through reports, seminars and its Blogs. Consequently, it would collaborate with the university’s departments on how to best utilize AI and IA to improve and make their teaching more exciting.