As recent years have progressed, technology has moved on and with it, the shadow of AI. AI was initially invented or discovered in 1956 when a workshop in the summer was arranged at Dartmouth College.
With the advent of supercomputers of even 10 years ago. The Cray T90 supercomputer of 1995 had a maximum of 57.6Gflps and modern workstations can have much more processing power than this out of the box, they also have the potential for more memory, meaning that AI is becoming more real every week.
The reasons why AI is getting bigger and bigger are multiple.
- Big data – data is produced in the digital economy at a massive rate. Storing it is still very expensive, and so is analysing it in some areas. Finance is a market that is data driven essentially and as such this feeds the AI algorithms and they improve performance over time.
- Moores law in effect – thus as mentioned previously, computers are so powerful in modern times that analysing big data has become cost effective.
- Human brain > into the algorithms brain: various hi tech algorithms eg speech recognition have continued to develop and are now at a useable level, whereas 5 years ago even they were prone to multiple innacuracies. Google even offers a free speech to text converter.
Investment – AI has not gone unnoticed in Silicon valley. And investments by big funds are commonplace in AI related startups. Not to mention the biggest players in the tech world like Google are buying up AI companies as fast as they can find them. They are buying not only the technology behind those companies, but also in essence the brainpower behind that tech.
1/ Asset management: AI algorithms will find connections between world news events and the effect asset prices. They can also monitor social media for trends and adjust asset investment accordingly.
2/ Credit scores: AI can help lenders make far more accurate and informed decisions on lending, and they can also add to this decision making process over time, from what they learned of correlations in the past.
3/ Detecting fraud: This has been used for years by credit card and banks, to detect fraudulent or suspicious spending or shopping activity.
AI can help a machine learn a clients spending pattern , and then monitor different accounts accordingly. For example, and AI algo in a bank will flag up a 90 year old ladies account (if she uses it very little and mainly for the odd cheque or direct debit) Vs a younger ‘yuppie’ type spender who is spending money far more frequently and in lots of different locations.
4/ AI market research: curates and then can index money market research and then even auto generate content to websites or other platforms using advanced text generation software.
5/ Customer service: AI customer support nodes can inspect incoming requests and direct them according to 1/priority /2 subject content /3 client or many other criteria. This use of the software cannot currently match real human interaction however going on the amount of complaints that AI staffed customer support desks get.