My Computer Just Had a Good Idea

Steven Sargent Artificial Intelligence, Blog, Technology 0 Comments

Creativity is the final bastion of what it is to be human, the one thing that machines are not going to be able to do, and therefore our jobs are safe provided they require creative input.  A new book out this month by Marcus du Sautoy, The Creativity Code, discusses to what extent computers can be programmed to invent and refine concepts.  Analogous to the well-known ‘Turing Test’ for machine intelligence, du Sautoy develops the concept of the ‘Lovelace Test’ for creativity.  Ada Lovelace was the Victorian computing visionary well known as the first computer programmer.  To pass the Lovelace test your computer must come up with something new, surprising and valuable and do so in a way that its programmers are not able to explain.

You could judge for yourself whether some recent examples of AI creativity would pass the test.  In 2017, the advertising agency Saatchi & Saatchi trained IBM’s supercomputer Watson to write thousands of items of advertising copy for Toyota’s Mirai car.

In 2018 a team of French entrepreneurs created original paintings resembling works by Old Masters such as Rembrandt using a computer algorithm.  Then they sold one of the paintings, “The Count of Belamy” for around £8,000 to a collector.

And in February this year, Elon-Musk-backed company OpenAI developed a language-processing algorithm described as “an extraordinary step forward, producing text rich with context, nuance and even something approaching humor” which is so good that OpenAI is not releasing its code in case it could be misused.

As the book explains Margaret Boden, an AI cognitive scientist now in her eighties, has classified creativity into three types; There is Exploratory creativity, Transformational creativity and Combination creativity. The last of which may well have the best potential for machine learning applications.  For this the combination needs to be something which can’t be logically deduced in advance from the separate parts.  An example from history is Robert Fulton’s development of the first commercially successful steamboat in 1807 by combining the steam engine and the sailing ship.

In the twenty-first century, here at 4C we are seeing how we can programme ‘Combination creativity’ into our procurement technology tools.  We call this using AI to develop ‘IA’ – ‘Inferencing Algorithms’.

Having information about the state of the supply market for your specialist category is all very well, but what if your procurement analysis software connected this together with your company spend trends and the contract compliance.  The ‘IA’ could then recommend that since your current contract was coming to an end, spend was increasing and there was enhanced supplier innovation, then you should invite tenders from newly identified suppliers for the predicted future spend amount.

Supplier risk assessment is a hot topic and is needed for both potential and contracted suppliers. What if your procurement analysis software provides supplier monitoring using syntax parsing to analyse natural language text into logical components and again connects into the spend and contracts databases.  The ‘IA’ would then flag up the most dangerous combinations of suppliers’ falling perception scores, specific category importance, increasing spend and out of contract, along with recommended courses of action.

Have you had your finance director asking you how your procurement department is contributing to the company EBITDA?  Using ‘IA’ and connecting periodic spend growth, spend under contract and the overall company revenue growth, you could report back on procurement’s contribution, and how you could increase this by having extra investment in the procurement function.

The final conclusion of the ‘Creativity Code’ seems to be that the most significant part of computer creativity comes from the creative input of the programming, something that we firmly believe in at 4C and that we continue to foster with the continuing developments in our procurement technology.

Leave a Reply

Your email address will not be published.